Daily Papers

Large language models (LLMs) often exhibit strong biases, e.g, against women or in favor of the number 7. We investigate whether LLMs would be able to output less biased answers when allowed to observe their prior answers to the same question in a multi-turn conversation. To understand which types of questions invite more biased answers, we test LLMs on our proposed set of questions that span 9 topics and belong to three types: (1) Subjective; (2) Random; and (3) Objective. Interestingly, LLMs are able to "de-bias" themselves in a multi-turn conversation in response to questions that seek an Random, unbiased answer. Furthermore, we propose B-score, a novel metric that is effective in detecting biases to Subjective, Random, Easy, and Hard questions. On MMLU, HLE, and CSQA, leveraging B-score substantially improves the verification accuracy of LLM answers (i.e, accepting LLM correct answers and rejecting incorrect ones) compared to using verbalized confidence scores or the frequency of single-turn answers alone. Code and data are available at: https://b-score.github.io.

We present our shared task on text-based emotion detection, covering more than 30 languages from seven distinct language families. These languages are predominantly low-resource and spoken across various continents. The data instances are multi-labeled into six emotional classes, with additional datasets in 11 languages annotated for emotion intensity. Participants were asked to predict labels in three tracks: (a) emotion labels in monolingual settings, (b) emotion intensity scores, and (c) emotion labels in cross-lingual settings. The task attracted over 700 participants. We received final submissions from more than 200 teams and 93 system description papers. We report baseline results, as well as findings on the best-performing systems, the most common approaches, and the most effective methods across various tracks and languages. The datasets for this task are publicly available.

Diffusion models are commonly interpreted as learning the score function, i.e., the gradient of the log-density of noisy data. However, this assumption implies that the target of learning is a conservative vector field, which is not enforced by the neural network architectures used in practice. We present numerical evidence that trained diffusion networks violate both integral and differential constraints required of true score functions, demonstrating that the learned vector fields are not conservative. Despite this, the models perform remarkably well as generative mechanisms. To explain this apparent paradox, we advocate a new theoretical perspective: diffusion training is better understood as flow matching to the velocity field of a Wasserstein Gradient Flow (WGF), rather than as score learning for a reverse-time stochastic differential equation. Under this view, the "probability flow" arises naturally from the WGF framework, eliminating the need to invoke reverse-time SDE theory and clarifying why generative sampling remains successful even when the neural vector field is not a true score. We further show that non-conservative errors from neural approximation do not necessarily harm density transport. Our results advocate for adopting the WGF perspective as a principled, elegant, and theoretically grounded framework for understanding diffusion generative models.

We introduce a novel framework for efficient sampling from complex, unnormalised target distributions by exploiting multiscale dynamics. Traditional score-based sampling methods either rely on learned approximations of the score function or involve computationally expensive nested Markov chain Monte Carlo (MCMC) loops. In contrast, the proposed approach leverages stochastic averaging within a slow-fast system of stochastic differential equations (SDEs) to estimate intermediate scores along a diffusion path without training or inner-loop MCMC. Two algorithms are developed under this framework: MultALMC, which uses multiscale annealed Langevin dynamics, and MultCDiff, based on multiscale controlled diffusions for the reverse-time Ornstein-Uhlenbeck process. Both overdamped and underdamped variants are considered, with theoretical guarantees of convergence to the desired diffusion path. The framework is extended to handle heavy-tailed target distributions using Student's t-based noise models and tailored fast-process dynamics. Empirical results across synthetic and real-world benchmarks, including multimodal and high-dimensional distributions, demonstrate that the proposed methods are competitive with existing samplers in terms of accuracy and efficiency, without the need for learned models.

Recent breakthroughs in text-to-4D generation rely on pre-trained text-to-image and text-to-video models to generate dynamic 3D scenes. However, current text-to-4D methods face a three-way tradeoff between the quality of scene appearance, 3D structure, and motion. For example, text-to-image models and their 3D-aware variants are trained on internet-scale image datasets and can be used to produce scenes with realistic appearance and 3D structure -- but no motion. Text-to-video models are trained on relatively smaller video datasets and can produce scenes with motion, but poorer appearance and 3D structure. While these models have complementary strengths, they also have opposing weaknesses, making it difficult to combine them in a way that alleviates this three-way tradeoff. Here, we introduce hybrid score distillation sampling, an alternating optimization procedure that blends supervision signals from multiple pre-trained diffusion models and incorporates benefits of each for high-fidelity text-to-4D generation. Using hybrid SDS, we demonstrate synthesis of 4D scenes with compelling appearance, 3D structure, and motion.

The soft Dice loss (SDL) has taken a pivotal role in numerous automated segmentation pipelines in the medical imaging community. Over the last years, some reasons behind its superior functioning have been uncovered and further optimizations have been explored. However, there is currently no implementation that supports its direct utilization in scenarios involving soft labels. Hence, a synergy between the use of SDL and research leveraging the use of soft labels, also in the context of model calibration, is still missing. In this work, we introduce Dice semimetric losses (DMLs), which (i) are by design identical to SDL in a standard setting with hard labels, but (ii) can be employed in settings with soft labels. Our experiments on the public QUBIQ, LiTS and KiTS benchmarks confirm the potential synergy of DMLs with soft labels (e.g.\ averaging, label smoothing, and knowledge distillation) over hard labels (e.g.\ majority voting and random selection). As a result, we obtain superior Dice scores and model calibration, which supports the wider adoption of DMLs in practice. The code is available at https://github.com/zifuwanggg/JDTLosses{https://github.com/zifuwanggg/JDTLosses}.

We propose a new policy representation based on score-based diffusion models (SDMs). We apply our new policy representation in the domain of Goal-Conditioned Imitation Learning (GCIL) to learn general-purpose goal-specified policies from large uncurated datasets without rewards. Our new goal-conditioned policy architecture "BEhavior generation with ScOre-based Diffusion Policies" (BESO) leverages a generative, score-based diffusion model as its policy. BESO decouples the learning of the score model from the inference sampling process, and, hence allows for fast sampling strategies to generate goal-specified behavior in just 3 denoising steps, compared to 30+ steps of other diffusion based policies. Furthermore, BESO is highly expressive and can effectively capture multi-modality present in the solution space of the play data. Unlike previous methods such as Latent Plans or C-Bet, BESO does not rely on complex hierarchical policies or additional clustering for effective goal-conditioned behavior learning. Finally, we show how BESO can even be used to learn a goal-independent policy from play-data using classifier-free guidance. To the best of our knowledge this is the first work that a) represents a behavior policy based on such a decoupled SDM b) learns an SDM based policy in the domain of GCIL and c) provides a way to simultaneously learn a goal-dependent and a goal-independent policy from play-data. We evaluate BESO through detailed simulation and show that it consistently outperforms several state-of-the-art goal-conditioned imitation learning methods on challenging benchmarks. We additionally provide extensive ablation studies and experiments to demonstrate the effectiveness of our method for goal-conditioned behavior generation. Demonstrations and Code are available at https://intuitive-robots.github.io/beso-website/

This paper shows that two commonly used evaluation metrics for generative models, the Fr\'echet Inception Distance (FID) and the Inception Score (IS), are biased -- the expected value of the score computed for a finite sample set is not the true value of the score. Worse, the paper shows that the bias term depends on the particular model being evaluated, so model A may get a better score than model B simply because model A's bias term is smaller. This effect cannot be fixed by evaluating at a fixed number of samples. This means all comparisons using FID or IS as currently computed are unreliable. We then show how to extrapolate the score to obtain an effectively bias-free estimate of scores computed with an infinite number of samples, which we term textrm{FID}_infty and textrm{IS}_infty. In turn, this effectively bias-free estimate requires good estimates of scores with a finite number of samples. We show that using Quasi-Monte Carlo integration notably improves estimates of FID and IS for finite sample sets. Our extrapolated scores are simple, drop-in replacements for the finite sample scores. Additionally, we show that using low discrepancy sequence in GAN training offers small improvements in the resulting generator.

Vision-language models (VLMs) discriminatively pre-trained with contrastive image-text matching losses such as P(match|text, image) have been criticized for lacking compositional understanding. This means they might output similar scores even if the original caption is rearranged into a different semantic statement. To address this, we propose to use the {bf V}isual {bf G}enerative {bf P}re-{bf T}raining Score ({bf VisualGPTScore}) of P(text|image), a multimodal generative score that captures the likelihood of a text caption conditioned on an image using an image-conditioned language model. Contrary to the belief that VLMs are mere bag-of-words models, our off-the-shelf VisualGPTScore demonstrates top-tier performance on recently proposed image-text retrieval benchmarks like ARO and Crepe that assess compositional reasoning. Furthermore, we factorize VisualGPTScore into a product of the marginal P(text) and the Pointwise Mutual Information (PMI). This helps to (a) diagnose datasets with strong language bias, and (b) debias results on other benchmarks like Winoground using an information-theoretic framework. VisualGPTScore provides valuable insights and serves as a strong baseline for future evaluation of visio-linguistic compositionality.

Multiple choice question answering (MCQA) is popular for LLM evaluation due to its simplicity and human-like testing, but we argue for its reform. We first reveal flaws in MCQA's format, as it struggles to: 1) test generation/subjectivity; 2) match LLM use cases; and 3) fully test knowledge. We instead advocate for generative formats based on human testing-where LLMs construct and explain answers-better capturing user needs and knowledge while remaining easy to score. We then show even when MCQA is a useful format, its datasets suffer from: leakage; unanswerability; shortcuts; and saturation. In each issue, we give fixes from education, like rubrics to guide MCQ writing; scoring methods to bridle guessing; and Item Response Theory to build harder MCQs. Lastly, we discuss LLM errors in MCQA-robustness, biases, and unfaithful explanations-showing how our prior solutions better measure or address these issues. While we do not need to desert MCQA, we encourage more efforts in refining the task based on educational testing, advancing evaluations.

In this work, we address the issue of quality of experience (QoE) in unmanned aerial vehicle (UAV) aided multiuser rate-splitting multiple access (RSMA) networks under secrecy constraints. The problem is formulated as maximization of sum mean opinion scores (MOSs) of the users. The problem is decomposed into two subproblems, beamforming and rate allocation and UAV trajectory subproblem. For, beamforming and rate allocation subproblem, we use the epigraph method, property of polynomials, and the norm-bounded error of channels, we linearize the objective function. Then, applying second-order conic (SOC) and first Taylor expansion, we convexify the remaining nonconvex constraints. For the highly nonconvex UAV trajectory, we unroll the constraints and we apply first Taylor expansion on the unrolled constraints. The simulation results demonstrate the efficiency of the proposed framework.

Kernelized Stein discrepancy (KSD) is a score-based discrepancy widely used in goodness-of-fit tests. It can be applied even when the target distribution has an unknown normalising factor, such as in Bayesian analysis. We show theoretically and empirically that the KSD test can suffer from low power when the target and the alternative distributions have the same well-separated modes but differ in mixing proportions. We propose to perturb the observed sample via Markov transition kernels, with respect to which the target distribution is invariant. This allows us to then employ the KSD test on the perturbed sample. We provide numerical evidence that with suitably chosen transition kernels the proposed approach can lead to substantially higher power than the KSD test.

Since their introduction, diffusion models have quickly become the prevailing approach to generative modeling in many domains. They can be interpreted as learning the gradients of a time-varying sequence of log-probability density functions. This interpretation has motivated classifier-based and classifier-free guidance as methods for post-hoc control of diffusion models. In this work, we build upon these ideas using the score-based interpretation of diffusion models, and explore alternative ways to condition, modify, and reuse diffusion models for tasks involving compositional generation and guidance. In particular, we investigate why certain types of composition fail using current techniques and present a number of solutions. We conclude that the sampler (not the model) is responsible for this failure and propose new samplers, inspired by MCMC, which enable successful compositional generation. Further, we propose an energy-based parameterization of diffusion models which enables the use of new compositional operators and more sophisticated, Metropolis-corrected samplers. Intriguingly we find these samplers lead to notable improvements in compositional generation across a wide set of problems such as classifier-guided ImageNet modeling and compositional text-to-image generation.

With 3.78 billion social media users worldwide in 2021 (48% of the human population), almost 3 billion images are shared daily. At the same time, a consistent evolution of smartphone cameras has led to a photography explosion with 85% of all new pictures being captured using smartphones. However, lately, there has been an increased discussion of privacy concerns when a person being photographed is unaware of the picture being taken or has reservations about the same being shared. These privacy violations are amplified for people with disabilities, who may find it challenging to raise dissent even if they are aware. Such unauthorized image captures may also be misused to gain sympathy by third-party organizations, leading to a privacy breach. Privacy for people with disabilities has so far received comparatively less attention from the AI community. This motivates us to work towards a solution to generate privacy-conscious cues for raising awareness in smartphone users of any sensitivity in their viewfinder content. To this end, we introduce PrivPAS (A real time Privacy-Preserving AI System) a novel framework to identify sensitive content. Additionally, we curate and annotate a dataset to identify and localize accessibility markers and classify whether an image is sensitive to a featured subject with a disability. We demonstrate that the proposed lightweight architecture, with a memory footprint of a mere 8.49MB, achieves a high mAP of 89.52% on resource-constrained devices. Furthermore, our pipeline, trained on face anonymized data, achieves an F1-score of 73.1%.

Sparse attention can reduce the cost of long-context inference, but most variants introduce new architectural components. We introduce Simplified Sparse Attention (SSA), a simpler approach to sparse attention that requires no architectural changes. Concretely, we first perform continued pretraining on sequences interleaved with gist tokens. We optimize the standard next-token loss as usual, but the gist tokens use an attention mask to restrict what parts of the context the language model can attend to; this teaches the model to pack each chunk's important information into the gist tokens. At inference time, SSA scores chunks via attention between the current query and the small set of gist tokens, selectively unfolding the top-k chunks by reintroducing their corresponding raw tokens. Since the query is scored only against the gist tokens, we avoid the memory-bandwidth cost associated with naive scoring against the full KV cache, without requiring the auxiliary KV cache approach used by sparse attention methods. On LongBench, SSA consistently outperforms compression and inference-time sparse-attention baselines under the same compression ratio. More strikingly, in retrieval-augmented generation, SSA can even outperform full attention after continued pretraining by over 5.7 points. We attribute this to the ability of SSA's selective unfolding, which concentrates attention on the query-relevant chunks and effectively filters out noise. SSA further extends to a hierarchical gist-of-gist variant (H-SSA) that achieves log-linear decoding complexity while maintaining or improving accuracy at high compression ratios up to 32x. The code is available at https://github.com/yuzhenmao/simplified-sparse-attention/.

A key barrier to interpreting large language models is polysemanticity, where neurons activate for multiple unrelated concepts. Sparse autoencoders (SAEs) have been proposed to mitigate this issue by transforming dense activations into sparse, more interpretable features. While prior work suggests that SAEs promote monosemanticity, no quantitative comparison has examined how concept activation distributions differ between SAEs and their base models. This paper provides the first systematic evaluation of SAEs against base models through activation distribution lens. We introduce a fine-grained concept separability score based on the Jensen-Shannon distance, which captures how distinctly a neuron's activation distributions vary across concepts. Using two large language models (Gemma-2-2B and DeepSeek-R1) and multiple SAE variants across five datasets (including word-level and sentence-level), we show that SAEs reduce polysemanticity and achieve higher concept separability. To assess practical utility, we evaluate concept-level interventions using two strategies: full neuron masking and partial suppression. We find that, compared to base models, SAEs enable more precise concept-level control when using partial suppression. Building on this, we propose Attenuation via Posterior Probabilities (APP), a new intervention method that uses concept-conditioned activation distributions for targeted suppression. APP achieves the smallest perplexity increase while remaining highly effective at concept removal.

Vision-language models can connect the text description of an object to its specific location in an image through visual grounding. This has potential applications in enhanced radiology reporting. However, these models require large annotated image-text datasets, which are lacking for PET/CT. We developed an automated pipeline to generate weak labels linking PET/CT report descriptions to their image locations and used it to train a 3D vision-language visual grounding model. Our pipeline finds positive findings in PET/CT reports by identifying mentions of SUVmax and axial slice numbers. From 25,578 PET/CT exams, we extracted 11,356 sentence-label pairs. Using this data, we trained ConTEXTual Net 3D, which integrates text embeddings from a large language model with a 3D nnU-Net via token-level cross-attention. The model's performance was compared against LLMSeg, a 2.5D version of ConTEXTual Net, and two nuclear medicine physicians. The weak-labeling pipeline accurately identified lesion locations in 98% of cases (246/251), with 7.5% requiring boundary adjustments. ConTEXTual Net 3D achieved an F1 score of 0.80, outperforming LLMSeg (F1=0.22) and the 2.5D model (F1=0.53), though it underperformed both physicians (F1=0.94 and 0.91). The model achieved better performance on FDG (F1=0.78) and DCFPyL (F1=0.75) exams, while performance dropped on DOTATE (F1=0.58) and Fluciclovine (F1=0.66). The model performed consistently across lesion sizes but showed reduced accuracy on lesions with low uptake. Our novel weak labeling pipeline accurately produced an annotated dataset of PET/CT image-text pairs, facilitating the development of 3D visual grounding models. ConTEXTual Net 3D significantly outperformed other models but fell short of the performance of nuclear medicine physicians. Our study suggests that even larger datasets may be needed to close this performance gap.

The rapid evolution of modern malware presents significant challenges to the development of effective defense mechanisms. Traditional cyber deception techniques often rely on static or manually configured parameters, limiting their adaptability to dynamic and sophisticated threats. This study leverages Generative AI (GenAI) models to automate the creation of adaptive cyber deception ploys, focusing on structured prompt engineering (PE) to enhance relevance, actionability, and deployability. We introduce a systematic framework (SPADE) to address inherent challenges large language models (LLMs) pose to adaptive deceptions, including generalized outputs, ambiguity, under-utilization of contextual information, and scalability constraints. Evaluations across diverse malware scenarios using metrics such as Recall, Exact Match (EM), BLEU Score, and expert quality assessments identified ChatGPT-4o as the top performer. Additionally, it achieved high engagement (93%) and accuracy (96%) with minimal refinements. Gemini and ChatGPT-4o Mini demonstrated competitive performance, with Llama3.2 showing promise despite requiring further optimization. These findings highlight the transformative potential of GenAI in automating scalable, adaptive deception strategies and underscore the critical role of structured PE in advancing real-world cybersecurity applications.

This investigation reports on the results of convolutional neural networks developed for the recently introduced PathologicAL Myopia (PALM) dataset, which consists of 1200 fundus images. We propose a new Optic Nerve Head (ONH)-based prediction enhancement for the segmentation of atrophy and fovea. Models trained with 400 available training images achieved an AUC of 0.9867 for pathological myopia classification, and a Euclidean distance of 58.27 pixels on the fovea localization task, evaluated on a test set of 400 images. Dice and F1 metrics for semantic segmentation of lesions scored 0.9303 and 0.9869 on optic disc, 0.8001 and 0.9135 on retinal atrophy, and 0.8073 and 0.7059 on retinal detachment, respectively. Our work was acknowledged with an award in the context of the "PathologicAL Myopia detection from retinal images" challenge held during the IEEE International Symposium on Biomedical Imaging (April 2019). Considering that (pathological) myopia cases are often identified as false positives and negatives in classification systems for glaucoma, we envision that the current work could aid in future research to discriminate between glaucomatous and highly-myopic eyes, complemented by the localization and segmentation of landmarks such as fovea, optic disc and atrophy.

Large text-to-video models trained on internet-scale data have demonstrated exceptional capabilities in generating high-fidelity videos from arbitrary textual descriptions. However, adapting these models to tasks with limited domain-specific data, such as animation or robotics videos, poses a significant computational challenge, since finetuning a pretrained large model can be prohibitively expensive. Inspired by how a small modifiable component (e.g., prompts, prefix-tuning) can adapt a large language model to perform new tasks without requiring access to the model weights, we investigate how to adapt a large pretrained text-to-video model to a variety of downstream domains and tasks without finetuning. In answering this question, we propose Video Adapter, which leverages the score function of a large pretrained video diffusion model as a probabilistic prior to guide the generation of a task-specific small video model. Our experiments show that Video Adapter is capable of incorporating the broad knowledge and preserving the high fidelity of a large pretrained video model in a task-specific small video model that is able to generate high-quality yet specialized videos on a variety of tasks such as animation, egocentric modeling, and modeling of simulated and real-world robotics data. More videos can be found on the website https://video-adapter.github.io/.

In material sciences, characterizing faults in periodic structures is vital for understanding material properties. To characterize magnetic labyrinthine patterns, it is necessary to accurately identify junctions and terminals, often featuring over a thousand closely packed defects per image. This study introduces a new technique called TM-CNN (Template Matching - Convolutional Neural Network) designed to detect a multitude of small objects in images, such as defects in magnetic labyrinthine patterns. TM-CNN was used to identify these structures in 444 experimental images, and the results were explored to deepen the understanding of magnetic materials. It employs a two-stage detection approach combining template matching, used in initial detection, with a convolutional neural network, used to eliminate incorrect identifications. To train a CNN classifier, it is necessary to create a large number of training images. This difficulty prevents the use of CNN in many practical applications. TM-CNN significantly reduces the manual workload for creating training images by automatically making most of the annotations and leaving only a small number of corrections to human reviewers. In testing, TM-CNN achieved an impressive F1 score of 0.988, far outperforming traditional template matching and CNN-based object detection algorithms.

Can pretrained models generalize to new datasets without any retraining? We deploy pretrained image models on datasets they were not trained for, and investigate whether their embeddings form meaningful clusters. Our suite of benchmarking experiments use encoders pretrained solely on ImageNet-1k with either supervised or self-supervised training techniques, deployed on image datasets that were not seen during training, and clustered with conventional clustering algorithms. This evaluation provides new insights into the embeddings of self-supervised models, which prioritize different features to supervised models. Supervised encoders typically offer more utility than SSL encoders within the training domain, and vice-versa far outside of it, however, fine-tuned encoders demonstrate the opposite trend. Clustering provides a way to evaluate the utility of self-supervised learned representations orthogonal to existing methods such as kNN. Additionally, we find the silhouette score when measured in a UMAP-reduced space is highly correlated with clustering performance, and can therefore be used as a proxy for clustering performance on data with no ground truth labels. Our code implementation is available at https://github.com/scottclowe/zs-ssl-clustering/.

Predicting gender by the first name is not a simple task. In many applications, especially in the natural language processing (NLP) field, this task may be necessary, mainly when considering foreign names. In this paper, we examined and implemented several machine learning algorithms, such as extra trees, KNN, Naive Bayes, SVM, random forest, gradient boosting, light GBM, logistic regression, ridge classifier, and deep neural network models, such as MLP, RNN, GRU, CNN, and BiLSTM, to classify gender through the first name. A dataset of Brazilian names is used to train and evaluate the models. We analyzed the accuracy, recall, precision, f1 score, and confusion matrix to measure the models' performances. The results indicate that the gender prediction can be performed from the feature extraction strategy looking at the names as a set of strings. Some models accurately predict gender in more than 95% of the cases. The recurrent models overcome the feedforward models in this binary classification problem.

Model-learning agents should gather information to learn world models that support many downstream tasks and inferences, such as predicting unobserved states, estimating near- and far-term consequences of actions, planning action sequences, and detecting changes in dynamics. Current methods for learning and evaluating world models diverge from this goal: training and evaluation are anchored to next-frame prediction, and success is scored by reward maximization in the same environment. We propose WorldTest, a protocol to evaluate model-learning agents that separates reward-free interaction from a scored test phase in a different but related environment. WorldTest is open-endedx2014models should support many different tasks unknown ahead of timex2014and agnostic to model representation, allowing comparison across approaches. We instantiated WorldTest with AutumnBench, a suite of 43 interactive grid-world environments and 129 tasks across three families: masked-frame prediction, planning, and predicting changes to the causal dynamics. We compared 517 human participants and three frontier models on AutumnBench. We found that humans outperform the models, and scaling compute improves performance only in some environments but not others. WorldTest provides a novel templatex2014reward-free exploration, derived tests, and behavior-based scoringx2014to evaluate what agents learn about environment dynamics, and AutumnBench exposes significant headroom in world-model learning.

Code-LLMs, LLMs pre-trained on large code corpora, have shown great progress in learning rich representations of the structure and syntax of code, successfully using it to generate or classify code fragments. At the same time, understanding if they are able to do so because they capture code semantics, and how well, is still an open question. In this paper, we tackle this problem by introducing SeqCoBench, a benchmark for systematically assessing how Code-LLMs can capture code functional equivalence. SeqCoBench contains over 20 code transformations that either preserve or alter the semantics of Python programs. We conduct extensive evaluations in different settings, including zero-shot and parameter-efficient finetuning methods on state-of-the-art (Code)-LLMs to see if they can discern semantically equivalent or different pairs of programs in SeqCoBench. We find that the performance gap between these LLMs and classical match-based retrieval scores is minimal, with both approaches showing a concerning lack of depth in understanding code semantics.

Audio-driven human animation technology is widely used in human-computer interaction, and the emergence of diffusion models has further advanced its development. Currently, most methods rely on multi-stage generation and intermediate representations, resulting in long inference time and issues with generation quality in specific foreground regions and audio-motion consistency. These shortcomings are primarily due to the lack of localized fine-grained supervised guidance. To address above challenges, we propose Parts-aware Audio-driven Human Animation, PAHA, a unit enhancement and guidance framework for audio-driven upper-body animation. We introduce two key methods: Parts-Aware Re-weighting (PAR) and Parts Consistency Enhancement (PCE). PAR dynamically adjusts regional training loss weights based on pose confidence scores, effectively improving visual quality. PCE constructs and trains diffusion-based regional audio-visual classifiers to improve the consistency of motion and co-speech audio. Afterwards, we design two novel inference guidance methods for the foregoing classifiers, Sequential Guidance (SG) and Differential Guidance (DG), to balance efficiency and quality respectively. Additionally, we build CNAS, the first public Chinese News Anchor Speech dataset, to advance research and validation in this field. Extensive experimental results and user studies demonstrate that PAHA significantly outperforms existing methods in audio-motion alignment and video-related evaluations. The codes and CNAS dataset will be released upon acceptance.

Super-resolution (SR), a classical inverse problem in computer vision, is inherently ill-posed, inducing a distribution of plausible solutions for every input. However, the desired result is not simply the expectation of this distribution, which is the blurry image obtained by minimizing pixelwise error, but rather the sample with the highest image quality. A variety of techniques, from perceptual metrics to adversarial losses, are employed to this end. In this work, we explore an alternative: utilizing powerful non-reference image quality assessment (NR-IQA) models in the SR context. We begin with a comprehensive analysis of NR-IQA metrics on human-derived SR data, identifying both the accuracy (human alignment) and complementarity of different metrics. Then, we explore two methods of applying NR-IQA models to SR learning: (i) altering data sampling, by building on an existing multi-ground-truth SR framework, and (ii) directly optimizing a differentiable quality score. Our results demonstrate a more human-centric perception-distortion tradeoff, focusing less on non-perceptual pixel-wise distortion, instead improving the balance between perceptual fidelity and human-tuned NR-IQA measures.

Existing large language model-based code generation pipelines typically use beam search or sampling algorithms during the decoding process. Although the programs they generate achieve high token-matching-based scores, they often fail to compile or generate incorrect outputs. The main reason is that conventional Transformer decoding algorithms may not be the best choice for code generation. In this work, we propose a novel Transformer decoding algorithm, Planning-Guided Transformer Decoding (PG-TD), that uses a planning algorithm to do lookahead search and guide the Transformer to generate better programs. Specifically, instead of simply optimizing the likelihood of the generated sequences, the Transformer makes use of a planner to generate candidate programs and test them on public test cases. The Transformer can therefore make more informed decisions and generate tokens that will eventually lead to higher-quality programs. We also design a mechanism that shares information between the Transformer and the planner to make our algorithm computationally efficient. We empirically evaluate our framework with several large language models as backbones on public coding challenge benchmarks, showing that 1) it can generate programs that consistently achieve higher performance compared with competing baseline methods; 2) it enables controllable code generation, such as concise codes and highly-commented codes by optimizing modified objective.

To what extent do vision-and-language foundation models possess a realistic world model (observation times action rightarrow observation) and a dynamics model (observation times observation rightarrow action), when actions are expressed through language? While open-source foundation models struggle with both, we find that fine-tuning them to acquire a dynamics model through supervision is significantly easier than acquiring a world model. In turn, dynamics models can be used to bootstrap world models through two main strategies: 1) weakly supervised learning from synthetic data and 2) inference time verification. Firstly, the dynamics model can annotate actions for unlabelled pairs of video frame observations to expand the training data. We further propose a new objective, where image tokens in observation pairs are weighted by their importance, as predicted by a recognition model. Secondly, the dynamics models can assign rewards to multiple samples of the world model to score them, effectively guiding search at inference time. We evaluate the world models resulting from both strategies through the task of action-centric image editing on Aurora-Bench. Our best model achieves a performance competitive with state-of-the-art image editing models, improving on them by a margin of 15% on real-world subsets according to GPT4o-as-judge, and achieving the best average human evaluation across all subsets of Aurora-Bench.

Rigorously evaluating machine intelligence against the broad spectrum of human general intelligence has become increasingly important and challenging in this era of rapid technological advance. Conventional AI benchmarks typically assess only narrow capabilities in a limited range of human activity. Most are also static, quickly saturating as developers explicitly or implicitly optimize for them. We propose that a more promising way to evaluate human-like general intelligence in AI systems is through a particularly strong form of general game playing: studying how and how well they play and learn to play all conceivable human games, in comparison to human players with the same level of experience, time, or other resources. We define a "human game" to be a game designed by humans for humans, and argue for the evaluative suitability of this space of all such games people can imagine and enjoy -- the "Multiverse of Human Games". Taking a first step towards this vision, we introduce the AI GameStore, a scalable and open-ended platform that uses LLMs with humans-in-the-loop to synthesize new representative human games, by automatically sourcing and adapting standardized and containerized variants of game environments from popular human digital gaming platforms. As a proof of concept, we generated 100 such games based on the top charts of Apple App Store and Steam, and evaluated seven frontier vision-language models (VLMs) on short episodes of play. The best models achieved less than 10\% of the human average score on the majority of the games, and especially struggled with games that challenge world-model learning, memory and planning. We conclude with a set of next steps for building out the AI GameStore as a practical way to measure and drive progress toward human-like general intelligence in machines.

The variable stars in the VSX catalog are derived from a multitude of inhomogeneous data sources and classification tools. This inhomogeneity complicates our understanding of variable star types, statistics, and properties, and it directly affects attempts to build training sets for current (and next) generation all-sky, time-domain surveys. We homogeneously analyze the ASAS-SN V-band light curves of {sim}412,000 variables from the VSX catalog. The variables are classified using an updated random forest classifier with an F_1 score of 99.4\% and refinement criteria for individual classifications. We have derived periods for {sim}52,000 variables in the VSX catalog that lack a period, and have reclassified {sim} 17,000 sources into new broad variability groups with high confidence. We have also reclassified {sim} 94,000 known variables with miscellaneous/generic classifications. The light curves, classifications, and a range of properties of the variables are all available through the ASAS-SN variable stars database (https://asas-sn.osu.edu/variables). We also include the V-band light curves for a set of {sim}4,000 rare variables and transient sources, including cataclysmic variables, symbiotic binaries and flare stars.

In this work, we report the set-up and results of the Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with the IEEE International Symposium on Biomedical Imaging (ISBI) 2017 and the International Conferences on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2017 and 2018. The image dataset is diverse and contains primary and secondary tumors with varied sizes and appearances with various lesion-to-background levels (hyper-/hypo-dense), created in collaboration with seven hospitals and research institutions. Seventy-five submitted liver and liver tumor segmentation algorithms were trained on a set of 131 computed tomography (CT) volumes and were tested on 70 unseen test images acquired from different patients. We found that not a single algorithm performed best for both liver and liver tumors in the three events. The best liver segmentation algorithm achieved a Dice score of 0.963, whereas, for tumor segmentation, the best algorithms achieved Dices scores of 0.674 (ISBI 2017), 0.702 (MICCAI 2017), and 0.739 (MICCAI 2018). Retrospectively, we performed additional analysis on liver tumor detection and revealed that not all top-performing segmentation algorithms worked well for tumor detection. The best liver tumor detection method achieved a lesion-wise recall of 0.458 (ISBI 2017), 0.515 (MICCAI 2017), and 0.554 (MICCAI 2018), indicating the need for further research. LiTS remains an active benchmark and resource for research, e.g., contributing the liver-related segmentation tasks in http://medicaldecathlon.com/. In addition, both data and online evaluation are accessible via www.lits-challenge.com.

Hit identification is a central challenge in early drug discovery, traditionally requiring substantial experimental resources. Recent advances in artificial intelligence, particularly large language models (LLMs), have enabled virtual screening methods that reduce costs and improve efficiency. However, the growing complexity of these tools has limited their accessibility to wet-lab researchers. Multi-agent systems offer a promising solution by combining the interpretability of LLMs with the precision of specialized models and tools. In this work, we present MADD, a multi-agent system that builds and executes customized hit identification pipelines from natural language queries. MADD employs four coordinated agents to handle key subtasks in de novo compound generation and screening. We evaluate MADD across seven drug discovery cases and demonstrate its superior performance compared to existing LLM-based solutions. Using MADD, we pioneer the application of AI-first drug design to five biological targets and release the identified hit molecules. Finally, we introduce a new benchmark of query-molecule pairs and docking scores for over three million compounds to contribute to the agentic future of drug design.

Attention steering is an important technique for controlling model focus, enabling capabilities such as prompt highlighting, where the model prioritises user-specified text. However, existing attention steering methods require explicit storage of the full attention matrix, making them incompatible with memory-efficient implementations like FlashAttention. We introduce Spectral Editing Key Amplification (SEKA), a training-free steering method that tackles this by directly editing key embeddings before attention computation. SEKA uses spectral decomposition to steer key embeddings towards latent directions that amplify attention scores for certain tokens. We extend this to Adaptive SEKA (AdaSEKA), a query-adaptive variant that uses a training-free routing mechanism to dynamically combine multiple expert subspaces based on the prompt's semantic intent. Our experiments show both methods significantly outperform strong baselines on standard steering benchmarks while adding much lower latency and memory overhead, in compatibility with optimised attention.

Logical reasoning, i.e., deductively inferring the truth value of a conclusion from a set of premises, is an important task for artificial intelligence with wide potential impacts on science, mathematics, and society. While many prompting-based strategies have been proposed to enable Large Language Models (LLMs) to do such reasoning more effectively, they still appear unsatisfactory, often failing in subtle and unpredictable ways. In this work, we investigate the validity of instead reformulating such tasks as modular neurosymbolic programming, which we call LINC: Logical Inference via Neurosymbolic Computation. In LINC, the LLM acts as a semantic parser, translating premises and conclusions from natural language to expressions in first-order logic. These expressions are then offloaded to an external theorem prover, which symbolically performs deductive inference. Leveraging this approach, we observe significant performance gains on FOLIO and a balanced subset of ProofWriter for three different models in nearly all experimental conditions we evaluate. On ProofWriter, augmenting the comparatively small open-source StarCoder+ (15.5B parameters) with LINC even outperforms GPT-3.5 and GPT-4 with Chain-of-Thought (CoT) prompting by an absolute 38% and 10%, respectively. When used with GPT-4, LINC scores 26% higher than CoT on ProofWriter while performing comparatively on FOLIO. Further analysis reveals that although both methods on average succeed roughly equally often on this dataset, they exhibit distinct and complementary failure modes. We thus provide promising evidence for how logical reasoning over natural language can be tackled through jointly leveraging LLMs alongside symbolic provers. All corresponding code is publicly available at https://github.com/benlipkin/linc

This paper introduces CoMuRoS (Collaborative Multi-Robot System), a generalizable hierarchical architecture for heterogeneous robot teams that unifies centralized deliberation with decentralized execution, and supports event-driven replanning. A Task Manager LLM interprets natural-language goals, classifies tasks, and allocates subtasks using static rules plus dynamic contexts (task, history, robot and task status, and events).Each robot runs a local LLM that composes executable Python code from primitive skills (ROS2 nodes, policies), while onboard perception (VLMs/image processing) continuously monitors events and classifies them into relevant or irrelevant to the task. Task failures or user intent changes trigger replanning, allowing robots to assist teammates, resume tasks, or request human help. Hardware studies demonstrate autonomous recovery from disruptive events, filtering of irrelevant distractions, and tightly coordinated transport with emergent human-robot cooperation (e.g., multirobot collaborative object recovery success rate: 9/10, coordinated transport: 8/8, human-assisted recovery: 5/5).Simulation studies show intention-aware replanning. A curated textual benchmark spanning 22 scenarios (3 tasks each, around 20 robots) evaluates task allocation, classification, IoU, executability, and correctness, with high average scores (e.g., correctness up to 0.91) across multiple LLMs, a separate replanning set (5 scenarios) achieves 1.0 correctness. Compared with prior LLM-based systems, CoMuRoS uniquely demonstrates runtime, event-driven replanning on physical robots, delivering robust, flexible multi-robot and human-robot collaboration.

African agriculture is undergoing rapid transformation. Annual maps of crop fields are key to understanding the nature of this transformation, but such maps are currently lacking and must be developed using advanced machine learning models trained on high resolution remote sensing imagery. To enable the development of such models, we delineated field boundaries in 33,746 Planet images captured between 2017 and 2023 across the continent using a custom labeling platform with built-in procedures for assessing and mitigating label error. We collected 42,403 labels, including 7,204 labels arising from tasks dedicated to assessing label quality (Class 1 labels), 32,167 from sites mapped once by a single labeller (Class 2) and 3,032 labels from sites where 3 or more labellers were tasked to map the same location (Class 4). Class 1 labels were used to calculate labeller-specific quality scores, while Class 1 and 4 sites mapped by at least 3 labellers were used to further evaluate label uncertainty using a Bayesian risk metric. Quality metrics showed that label quality was moderately high (0.75) for measures of total field extent, but low regarding the number of individual fields delineated (0.33), and the position of field edges (0.05). These values are expected when delineating small-scale fields in 3-5 m resolution imagery, which can be too coarse to reliably distinguish smaller fields, particularly in dense croplands, and therefore requires substantial labeller judgement. Nevertheless, previous work shows that such labels can train effective field mapping models. Furthermore, this large, probabilistic sample on its own provides valuable insight into regional agricultural characteristics, highlighting variations in the median field size and density. The imagery and vectorized labels along with quality information is available for download from two public repositories.

Action Quality Assessment (AQA) evaluates diverse skills but models struggle with non-stationary data. We propose Continual AQA (CAQA) to refine models using sparse new data. Feature replay preserves memory without storing raw inputs. However, the misalignment between static old features and the dynamically changing feature manifold causes severe catastrophic forgetting. To address this novel problem, we propose Manifold-Aligned Graph Regularization (MAGR), which first aligns deviated old features to the current feature manifold, ensuring representation consistency. It then constructs a graph jointly arranging old and new features aligned with quality scores. Experiments show MAGR outperforms recent strong baselines with up to 6.56%, 5.66%, 15.64%, and 9.05% correlation gains on the MTL-AQA, FineDiving, UNLV-Dive, and JDM-MSA split datasets, respectively. This validates MAGR for continual assessment challenges arising from non-stationary skill variations.

We introduce iterative reasoning through energy diffusion (IRED), a novel framework for learning to reason for a variety of tasks by formulating reasoning and decision-making problems with energy-based optimization. IRED learns energy functions to represent the constraints between input conditions and desired outputs. After training, IRED adapts the number of optimization steps during inference based on problem difficulty, enabling it to solve problems outside its training distribution -- such as more complex Sudoku puzzles, matrix completion with large value magnitudes, and pathfinding in larger graphs. Key to our method's success is two novel techniques: learning a sequence of annealed energy landscapes for easier inference and a combination of score function and energy landscape supervision for faster and more stable training. Our experiments show that IRED outperforms existing methods in continuous-space reasoning, discrete-space reasoning, and planning tasks, particularly in more challenging scenarios. Code and visualizations at https://energy-based-model.github.io/ired/

This paper provides a review of the NTIRE 2026 challenge on real-world face restoration, highlighting the proposed solutions and the resulting outcomes. The challenge focuses on generating natural and realistic outputs while maintaining identity consistency. Its goal is to advance state-of-the-art solutions for perceptual quality and realism, without imposing constraints on computational resources or training data. Performance is evaluated using a weighted image quality assessment (IQA) score and employs the AdaFace model as an identity checker. The competition attracted 96 registrants, with 10 teams submitting valid models; ultimately, 9 teams achieved valid scores in the final ranking. This collaborative effort advances the performance of real-world face restoration while offering an in-depth overview of the latest trends in the field.

We propose a unified framework for Singing Voice Synthesis (SVS) and Conversion (SVC), addressing the limitations of existing approaches in cross-domain SVS/SVC, poor output musicality, and scarcity of singing data. Our framework enables control over multiple aspects, including language content based on lyrics, performance attributes based on a musical score, singing style and vocal techniques based on a selector, and voice identity based on a speech sample. The proposed zero-shot learning paradigm consists of one SVS model and two SVC models, utilizing pre-trained content embeddings and a diffusion-based generator. The proposed framework is also trained on mixed datasets comprising both singing and speech audio, allowing singing voice cloning based on speech reference. Experiments show substantial improvements in timbre similarity and musicality over state-of-the-art baselines, providing insights into other low-data music tasks such as instrumental style transfer. Examples can be found at: everyone-can-sing.github.io.

The Philippine Government recently passed the "National Writing System Act," which promotes using Baybayin in Philippine texts. In support of this effort to promote the use of Baybayin, we present a computer vision system which can aid individuals who cannot easily read Baybayin script. In this paper, we survey the existing methods of identifying Baybayin scripts using computer vision and machine learning techniques and discuss their capabilities and limitations. Further, we propose a Baybayin Optical Character Instance Segmentation and Classification model using state-of-the-art Convolutional Neural Networks (CNNs) that detect Baybayin character instances in an image then outputs the Latin alphabet counterparts of each character instance in the image. Most existing systems are limited to character-level image classification and often misclassify or not natively support characters with diacritics. In addition, these existing models often have specific input requirements that limit it to classifying Baybayin text in a controlled setting, such as limitations in clarity and contrast, among others. To our knowledge, our proposed method is the first end-to-end character instance detection model for Baybayin, achieving a mAP50 score of 93.30%, mAP50-95 score of 80.50%, and F1-Score of 84.84%.

This paper tackles the problem of positioning a swarm of UAVs inside a completely unknown terrain, having as objective to maximize the overall situational awareness. The situational awareness is expressed by the number and quality of unique objects of interest, inside the UAVs' fields of view. YOLOv3 and a system to identify duplicate objects of interest were employed to assign a single score to each UAVs' configuration. Then, a novel navigation algorithm, capable of optimizing the previously defined score, without taking into consideration the dynamics of either UAVs or environment, is proposed. A cornerstone of the proposed approach is that it shares the same convergence characteristics as the block coordinate descent (BCD) family of approaches. The effectiveness and performance of the proposed navigation scheme were evaluated utilizing a series of experiments inside the AirSim simulator. The experimental evaluation indicates that the proposed navigation algorithm was able to consistently navigate the swarm of UAVs to "strategic" monitoring positions and also adapt to the different number of swarm sizes. Source code is available at https://github.com/dimikout3/ConvCAOAirSim.

Deep convolutional artificial neural networks (ANNs) are the leading class of candidate models of the mechanisms of visual processing in the primate ventral stream. While initially inspired by brain anatomy, over the past years, these ANNs have evolved from a simple eight-layer architecture in AlexNet to extremely deep and branching architectures, demonstrating increasingly better object categorization performance, yet bringing into question how brain-like they still are. In particular, typical deep models from the machine learning community are often hard to map onto the brain's anatomy due to their vast number of layers and missing biologically-important connections, such as recurrence. Here we demonstrate that better anatomical alignment to the brain and high performance on machine learning as well as neuroscience measures do not have to be in contradiction. We developed CORnet-S, a shallow ANN with four anatomically mapped areas and recurrent connectivity, guided by Brain-Score, a new large-scale composite of neural and behavioral benchmarks for quantifying the functional fidelity of models of the primate ventral visual stream. Despite being significantly shallower than most models, CORnet-S is the top model on Brain-Score and outperforms similarly compact models on ImageNet. Moreover, our extensive analyses of CORnet-S circuitry variants reveal that recurrence is the main predictive factor of both Brain-Score and ImageNet top-1 performance. Finally, we report that the temporal evolution of the CORnet-S "IT" neural population resembles the actual monkey IT population dynamics. Taken together, these results establish CORnet-S, a compact, recurrent ANN, as the current best model of the primate ventral visual stream.

Large Language Models (LLMs) present immense potential in the medical field, yet concerns over data privacy, regulatory compliance, and model stability restrict their widespread adoption. Although the distillation of high-performing closed-source LLMs has proven effective for general tasks, their application in healthcare is limited due to reduced domain knowledge and remnants of alignment behavior hindering clinical tasks. To address these challenges, we propose Dialogue-Based Knowledge Encoding (DBKE). DBKE enhances models' implicit knowledge base and primes them for conversational recall, augmenting their conversational capabilities and enabling a soft alignment for subsequent use cases. By transforming dense academic source text into synthetic dialogue, DBKE broadens the model's knowledge base and enables a soft alignment that guides downstream behaviours. We present Clinical Camel, an open-source, healthcare-focused conversational model, to showcase the effectiveness of DBKE. Clinical Camel outperforms GPT-3.5 on the United States Medical Licensing Examination (USMLE) Step 1 and Step 3 with scores of 53.2 % and 58.2 %, respectively, compared to GPT-3.5's scores of 36.1 % and 55.7 %. Clinical Camel adeptly handles multi-stage clinical case problems, provides adaptive counseling, and generates clinical notes. However, it is prone to hallucinations, which pose a significant obstacle in safety-critical settings. The performance of Clinical Camel underscores the importance of continued research and development of open-source models for the safe and effective integration of LLMs in healthcare settings.

Audio-visual understanding is a rapidly evolving field that seeks to integrate and interpret information from both auditory and visual modalities. Despite recent advances in multi-modal learning, existing benchmarks often suffer from strong visual bias -- where answers can be inferred from visual data alone -- and provide only aggregate scores that conflate multiple sources of error. This makes it difficult to determine whether models struggle with visual understanding, audio interpretation, or audio-visual alignment. In this work, we introduce DAVE (Diagnostic Audio Visual Evaluation), a novel benchmark dataset designed to systematically evaluate audio-visual models across controlled challenges. DAVE alleviates existing limitations by (i) ensuring both modalities are necessary to answer correctly and (ii) decoupling evaluation into atomic subcategories. Our detailed analysis of state-of-the-art models reveals specific failure modes and provides targeted insights for improvement. By offering this standardized diagnostic framework, we aim to facilitate more robust development of audio-visual models. The dataset is released: https://github.com/gorjanradevski/dave

We present OrbNet Denali, a machine learning model for electronic structure that is designed as a drop-in replacement for ground-state density functional theory (DFT) energy calculations. The model is a message-passing neural network that uses symmetry-adapted atomic orbital features from a low-cost quantum calculation to predict the energy of a molecule. OrbNet Denali is trained on a vast dataset of 2.3 million DFT calculations on molecules and geometries. This dataset covers the most common elements in bio- and organic chemistry (H, Li, B, C, N, O, F, Na, Mg, Si, P, S, Cl, K, Ca, Br, I) as well as charged molecules. OrbNet Denali is demonstrated on several well-established benchmark datasets, and we find that it provides accuracy that is on par with modern DFT methods while offering a speedup of up to three orders of magnitude. For the GMTKN55 benchmark set, OrbNet Denali achieves WTMAD-1 and WTMAD-2 scores of 7.19 and 9.84, on par with modern DFT functionals. For several GMTKN55 subsets, which contain chemical problems that are not present in the training set, OrbNet Denali produces a mean absolute error comparable to those of DFT methods. For the Hutchison conformers benchmark set, OrbNet Denali has a median correlation coefficient of R^2=0.90 compared to the reference DLPNO-CCSD(T) calculation, and R^2=0.97 compared to the method used to generate the training data (wB97X-D3/def2-TZVP), exceeding the performance of any other method with a similar cost. Similarly, the model reaches chemical accuracy for non-covalent interactions in the S66x10 dataset. For torsional profiles, OrbNet Denali reproduces the torsion profiles of wB97X-D3/def2-TZVP with an average MAE of 0.12 kcal/mol for the potential energy surfaces of the diverse fragments in the TorsionNet500 dataset.

This paper presents a stochastic differential equation (SDE) approach for general-purpose image restoration. The key construction consists in a mean-reverting SDE that transforms a high-quality image into a degraded counterpart as a mean state with fixed Gaussian noise. Then, by simulating the corresponding reverse-time SDE, we are able to restore the origin of the low-quality image without relying on any task-specific prior knowledge. Crucially, the proposed mean-reverting SDE has a closed-form solution, allowing us to compute the ground truth time-dependent score and learn it with a neural network. Moreover, we propose a maximum likelihood objective to learn an optimal reverse trajectory that stabilizes the training and improves the restoration results. The experiments show that our proposed method achieves highly competitive performance in quantitative comparisons on image deraining, deblurring, and denoising, setting a new state-of-the-art on two deraining datasets. Finally, the general applicability of our approach is further demonstrated via qualitative results on image super-resolution, inpainting, and dehazing. Code is available at https://github.com/Algolzw/image-restoration-sde.

Molecule generation is advancing rapidly in chemical discovery and drug design. Flow matching methods have recently set the state of the art (SOTA) in unconditional molecule generation, surpassing score-based diffusion models. However, diffusion models still lead in property-guided generation. In this work, we introduce PropMolFlow, a novel approach for property-guided molecule generation based on geometry-complete SE(3)-equivariant flow matching. Integrating five different property embedding methods with a Gaussian expansion of scalar properties, PropMolFlow outperforms previous SOTA diffusion models in conditional molecule generation across various properties while preserving the stability and validity of the generated molecules, consistent with its unconditional counterpart. Additionally, it enables faster inference with significantly fewer time steps compared to baseline models. We highlight the importance of validating the properties of generated molecules through DFT calculations performed at the same level of theory as the training data. Specifically, our analysis identifies properties that require DFT validation and others where a pretrained SE(3) geometric vector perceptron regressors provide sufficiently accurate predictions on generated molecules. Furthermore, we introduce a new property metric designed to assess the model's ability to propose molecules with underrepresented property values, assessing its capacity for out-of-distribution generalization. Our findings reveal shortcomings in existing structural metrics, which mistakenly validate open-shell molecules or molecules with invalid valence-charge configurations, underscoring the need for improved evaluation frameworks. Overall, this work paves the way for developing targeted property-guided generation methods, enhancing the design of molecular generative models for diverse applications.

It has been shown that dual encoders trained on one domain often fail to generalize to other domains for retrieval tasks. One widespread belief is that the bottleneck layer of a dual encoder, where the final score is simply a dot-product between a query vector and a passage vector, is too limited to make dual encoders an effective retrieval model for out-of-domain generalization. In this paper, we challenge this belief by scaling up the size of the dual encoder model {\em while keeping the bottleneck embedding size fixed.} With multi-stage training, surprisingly, scaling up the model size brings significant improvement on a variety of retrieval tasks, especially for out-of-domain generalization. Experimental results show that our dual encoders, Generalizable T5-based dense Retrievers (GTR), outperform %ColBERT~khattab2020colbert and existing sparse and dense retrievers on the BEIR dataset~thakur2021beir significantly. Most surprisingly, our ablation study finds that GTR is very data efficient, as it only needs 10\% of MS Marco supervised data to achieve the best out-of-domain performance. All the GTR models are released at https://tfhub.dev/google/collections/gtr/1.

Reward learning enables the application of reinforcement learning (RL) to tasks where reward is defined by human judgment, building a model of reward by asking humans questions. Most work on reward learning has used simulated environments, but complex information about values is often expressed in natural language, and we believe reward learning for language is a key to making RL practical and safe for real-world tasks. In this paper, we build on advances in generative pretraining of language models to apply reward learning to four natural language tasks: continuing text with positive sentiment or physically descriptive language, and summarization tasks on the TL;DR and CNN/Daily Mail datasets. For stylistic continuation we achieve good results with only 5,000 comparisons evaluated by humans. For summarization, models trained with 60,000 comparisons copy whole sentences from the input but skip irrelevant preamble; this leads to reasonable ROUGE scores and very good performance according to our human labelers, but may be exploiting the fact that labelers rely on simple heuristics.

Best-of-n sampling improves the accuracy of large language models (LLMs) and large reasoning models (LRMs) by generating multiple candidate solutions and selecting the one with the highest reward. The key challenge for reasoning tasks is designing a scoring function that can identify correct reasoning chains without access to ground-truth answers. We propose Probabilistic Confidence Selection And Ranking (PiCSAR): a simple, training-free method that scores each candidate generation using the joint log-likelihood of the reasoning and final answer. The joint log-likelihood of the reasoning and final answer naturally decomposes into reasoning confidence and answer confidence. PiCSAR achieves substantial gains across diverse benchmarks (+10.18 on MATH500, +9.81 on AIME2025), outperforming baselines with at least 2x fewer samples in 16 out of 20 comparisons. Our analysis reveals that correct reasoning chains exhibit significantly higher reasoning and answer confidence, justifying the effectiveness of PiCSAR.

Physicians now pose millions of clinical questions to AI tools each week, yet these tools are evaluated largely on hypothetical or exam-style questions, not those actually asked in practice. We report a blinded evaluation built on 620 Real-world Point-Of-Care Queries (Real-POCQi) submitted to the OpenEvidence (OE) platform by physicians spanning 30 specialties, as well as 187 questions from HealthBench. 149 practicing physicians across 36 states made head-to-head comparisons between answers from three frontier general-purpose models (Claude Opus 4.8, Gemini 3.1 Pro, and GPT-5.5) and a specialized clinical tool (OE), with graders matched to each question's specialty. When comparing answers along five dimensions relevant to clinical decision support -- accuracy, clinical utility, source quality, verifiability, & completeness -- physicians scored the specialized tool highest on all axes; in the primary analysis on Real-POCQi, win differences (margins between win and loss rates) ranged from 25 to 39 percentage points (p<0.001). Results remained consistent in sensitivity analyses stratifying by citation display, answer length, OE-user status, and Real-POCQi versus HealthBench. In parallel, LLM judges were found to systematically differ from expert judges, though both generally agreed on the best model. These findings underscore two conclusions: (i) AI tool evaluations should reflect real-world query distributions and use expert judges that mirror the specialization defining modern medicine and (ii) the consistent advantage of the specialized tool over general-purpose models does not necessarily mean that the latter cannot serve similar purposes, but that targeted engineering and customization can yield meaningful gains in performance for its users. We release Real-POCQi as a public benchmark, as well as the prespecified statistical analysis for reproducing results of this study.

Counting and tracking dense crowds in large-scale scenes is a highly practical yet challenging problem. Existing methods mostly rely on fixed-camera datasets with limited scene coverage, making them inadequate for crowd analysis in large-scale scenes. To bridge this gap, we introduce MovingDroneCrowd++, the largest video-level dataset dedicated to dense crowd counting and tracking with fast-moving drones, captured under diverse flight altitudes, camera angles, and illumination conditions. Existing methods, however, still fail to achieve satisfactory video individual counting or tracking performance under these challenging aerial conditions. To this end, we propose GD3A (Global Density map Decomposition via group-wise Descriptor Association), a video individual counting method that first establishes pixel-level correspondences between pedestrian descriptors across frames via optimal transport with an adaptive dustbin score. Then, group-wise association is adopted to guide the decomposition of the global density map into shared, inflow, and outflow density maps. We further introduce a pedestrian tracking method, DVTrack (Descriptor Voting Track), which converts descriptor-level matching into instance-level association through descriptor voting. Our methods rely on the association results of group-wise multiple descriptors for each pedestrian rather than a single vector. Since intra-group matching errors do not affect the final counting and tracking results, our methods are more robust in dense crowds and challenging aerial conditions. Experiments show that our methods achieve substantial gains in both crowd counting and tracking on moving-drone videos with dense crowds and complex motions, reducing counting error by 47.4% and improving tracking accuracy by 64.6%. Code, dataset, and pretrained models are available at https://github.com/fyw1999/MovingDroneCrowd.

Hospitals struggle to predict critical outcomes. Traditional early warning systems, like NEWS and MEWS, rely on static variables and fixed thresholds, limiting their adaptability, accuracy, and personalization. We previously developed the Enhanced Transformer for Health Outcome Simulation (ETHOS), an AI model that tokenizes patient health timelines (PHTs) from EHRs and uses transformer-based architectures to predict future PHTs. ETHOS is a versatile framework for developing a wide range of applications. In this work, we develop the Adaptive Risk Estimation System (ARES) that leverages ETHOS to compute dynamic, personalized risk probabilities for clinician-defined critical events. ARES also features a personalized explainability module that highlights key clinical factors influencing risk estimates. We evaluated ARES using the MIMIC-IV v2.2 dataset together with its Emergency Department (ED) extension and benchmarked performance against both classical early warning systems and contemporary machine learning models. The entire dataset was tokenized resulting in 285,622 PHTs, comprising over 360 million tokens. ETHOS outperformed benchmark models in predicting hospital admissions, ICU admissions, and prolonged stays, achieving superior AUC scores. Its risk estimates were robust across demographic subgroups, with calibration curves confirming model reliability. The explainability module provided valuable insights into patient-specific risk factors. ARES, powered by ETHOS, advances predictive healthcare AI by delivering dynamic, real-time, personalized risk estimation with patient-specific explainability. Although our results are promising, the clinical impact remains uncertain. Demonstrating ARES's true utility in real-world settings will be the focus of our future work. We release the source code to facilitate future research.

Detecting online hate is a difficult task that even state-of-the-art models struggle with. Typically, hate speech detection models are evaluated by measuring their performance on held-out test data using metrics such as accuracy and F1 score. However, this approach makes it difficult to identify specific model weak points. It also risks overestimating generalisable model performance due to increasingly well-evidenced systematic gaps and biases in hate speech datasets. To enable more targeted diagnostic insights, we introduce HateCheck, a suite of functional tests for hate speech detection models. We specify 29 model functionalities motivated by a review of previous research and a series of interviews with civil society stakeholders. We craft test cases for each functionality and validate their quality through a structured annotation process. To illustrate HateCheck's utility, we test near-state-of-the-art transformer models as well as two popular commercial models, revealing critical model weaknesses.

Tuberculosis (TB) is a chronic lung disease that occurs due to bacterial infection and is one of the top 10 leading causes of death. Accurate and early detection of TB is very important, otherwise, it could be life-threatening. In this work, we have detected TB reliably from the chest X-ray images using image pre-processing, data augmentation, image segmentation, and deep-learning classification techniques. Several public databases were used to create a database of 700 TB infected and 3500 normal chest X-ray images for this study. Nine different deep CNNs (ResNet18, ResNet50, ResNet101, ChexNet, InceptionV3, Vgg19, DenseNet201, SqueezeNet, and MobileNet), which were used for transfer learning from their pre-trained initial weights and trained, validated and tested for classifying TB and non-TB normal cases. Three different experiments were carried out in this work: segmentation of X-ray images using two different U-net models, classification using X-ray images, and segmented lung images. The accuracy, precision, sensitivity, F1-score, specificity in the detection of tuberculosis using X-ray images were 97.07 %, 97.34 %, 97.07 %, 97.14 % and 97.36 % respectively. However, segmented lungs for the classification outperformed than whole X-ray image-based classification and accuracy, precision, sensitivity, F1-score, specificity were 99.9 %, 99.91 %, 99.9 %, 99.9 %, and 99.52 % respectively. The paper also used a visualization technique to confirm that CNN learns dominantly from the segmented lung regions results in higher detection accuracy. The proposed method with state-of-the-art performance can be useful in the computer-aided faster diagnosis of tuberculosis.

Neurological diseases are the leading global cause of disability, yet most lack disease-modifying treatments. We present PROTON, a heterogeneous graph transformer that generates testable hypotheses across molecular, organoid, and clinical systems. To evaluate PROTON, we apply it to Parkinson's disease (PD), bipolar disorder (BD), and Alzheimer's disease (AD). In PD, PROTON linked genetic risk loci to genes essential for dopaminergic neuron survival and predicted pesticides toxic to patient-derived neurons, including the insecticide endosulfan, which ranked within the top 1.29% of predictions. In silico screens performed by PROTON reproduced six genome-wide α-synuclein experiments, including a split-ubiquitin yeast two-hybrid system (normalized enrichment score [NES] = 2.30, FDR-adjusted p < 1 times 10^{-4}), an ascorbate peroxidase proximity labeling assay (NES = 2.16, FDR < 1 times 10^{-4}), and a high-depth targeted exome sequencing study in 496 synucleinopathy patients (NES = 2.13, FDR < 1 times 10^{-4}). In BD, PROTON predicted calcitriol as a candidate drug that reversed proteomic alterations observed in cortical organoids derived from BD patients. In AD, we evaluated PROTON predictions in health records from n = 610,524 patients at Mass General Brigham, confirming that five PROTON-predicted drugs were associated with reduced seven-year dementia risk (minimum hazard ratio = 0.63, 95% CI: 0.53-0.75, p < 1 times 10^{-7}). PROTON generated neurological hypotheses that were evaluated across molecular, organoid, and clinical systems, defining a path for AI-driven discovery in neurological disease.

Automated singing assessment is crucial for education and entertainment. However, existing systems face two fundamental limitations: reliance on reference tracks, which stifles creative expression, and the simplification of complex performances into non-diagnostic scores based solely on pitch and rhythm. We advocate for a shift from discriminative to descriptive evaluation, creating a complete ecosystem for reference-free, multi-dimensional assessment. First, we introduce Sing-MD, a large-scale dataset annotated by experts across four dimensions: breath control, timbre quality, emotional expression, and vocal technique. Our analysis reveals significant annotation inconsistencies among experts, challenging the validity of traditional accuracy-based metrics. Second, addressing the memory limitations of Multimodal Large Language Models (MLLMs) in analyzing full-length songs, we propose VocalVerse. This efficient hybrid architecture leverages a lightweight acoustic encoder to model global performance features and long-term dependencies. Third, to address automated metric shortcomings, we establish the H-TPR (Human-in-the-loop Tiered Perceptual Ranking) benchmark, which evaluates a model's ability to generate perceptually valid rankings rather than predicting noisy ground-truth scores.

With the growing use of language models (LMs) in clinical environments, there is an immediate need to evaluate the accuracy and safety of LM-generated medical text. Currently, such evaluation relies solely on manual physician review. However, detecting errors in LM-generated text is challenging because 1) manual review is costly and 2) expert-composed reference outputs are often unavailable in real-world settings. While the "LM-as-judge" paradigm (a LM evaluating another LM) offers scalable evaluation, even frontier LMs can miss subtle but clinically significant errors. To address these challenges, we propose MedVAL, a self-supervised framework that leverages synthetic data to train evaluator LMs to assess whether LM-generated medical outputs are factually consistent with inputs, without requiring physician labels or reference outputs. To evaluate LM performance, we introduce MedVAL-Bench, a dataset containing 840 outputs annotated by physicians, following a physician-defined taxonomy of risk levels and error categories. Across 6 diverse medical tasks and 10 state-of-the-art LMs spanning open-source, proprietary, and medically adapted models, MedVAL fine-tuning significantly improves (p < 0.001) alignment with physicians on both seen and unseen tasks, increasing average F1 scores from 66% to 83%, with per-sample safety classification scores up to 86%. MedVAL improves the performance of even the best-performing proprietary LM (GPT-4o) by 8%. To support a scalable, risk-aware pathway towards clinical integration, we open-source the 1) codebase ( https://github.com/StanfordMIMI/MedVAL ), 2) MedVAL-Bench ( https://huggingface.co/datasets/stanfordmimi/MedVAL-Bench ), and 3) MedVAL-4B ( https://huggingface.co/stanfordmimi/MedVAL-4B ), the best-performing open-source LM. Our research provides the first evidence of LMs approaching expert-level validation ability for medical text.

A key component of deep learning (DL) for natural language processing (NLP) is word embeddings. Word embeddings that effectively capture the meaning and context of the word that they represent can significantly improve the performance of downstream DL models for various NLP tasks. Many existing word embeddings techniques capture the context of words based on word co-occurrence in documents and text; however, they often cannot capture broader domain-specific relationships between concepts that may be crucial for the NLP task at hand. In this paper, we propose a method to integrate external knowledge from medical terminology ontologies into the context captured by word embeddings. Specifically, we use a medical knowledge graph, such as the unified medical language system (UMLS), to find connections between clinical terms in cancer pathology reports. This approach aims to minimize the distance between connected clinical concepts. We evaluate the proposed approach using a Multitask Convolutional Neural Network (MT-CNN) to extract six cancer characteristics -- site, subsite, laterality, behavior, histology, and grade -- from a dataset of ~900K cancer pathology reports. The results show that the MT-CNN model which uses our domain informed embeddings outperforms the same MT-CNN using standard word2vec embeddings across all tasks, with an improvement in the overall micro- and macro-F1 scores by 4.97\%and 22.5\%, respectively.

We present the findings of "The Alzheimer's Disease Prediction Of Longitudinal Evolution" (TADPOLE) Challenge, which compared the performance of 92 algorithms from 33 international teams at predicting the future trajectory of 219 individuals at risk of Alzheimer's disease. Challenge participants were required to make a prediction, for each month of a 5-year future time period, of three key outcomes: clinical diagnosis, Alzheimer's Disease Assessment Scale Cognitive Subdomain (ADAS-Cog13), and total volume of the ventricles. The methods used by challenge participants included multivariate linear regression, machine learning methods such as support vector machines and deep neural networks, as well as disease progression models. No single submission was best at predicting all three outcomes. For clinical diagnosis and ventricle volume prediction, the best algorithms strongly outperform simple baselines in predictive ability. However, for ADAS-Cog13 no single submitted prediction method was significantly better than random guesswork. Two ensemble methods based on taking the mean and median over all predictions, obtained top scores on almost all tasks. Better than average performance at diagnosis prediction was generally associated with the additional inclusion of features from cerebrospinal fluid (CSF) samples and diffusion tensor imaging (DTI). On the other hand, better performance at ventricle volume prediction was associated with inclusion of summary statistics, such as the slope or maxima/minima of biomarkers. TADPOLE's unique results suggest that current prediction algorithms provide sufficient accuracy to exploit biomarkers related to clinical diagnosis and ventricle volume, for cohort refinement in clinical trials for Alzheimer's disease. However, results call into question the usage of cognitive test scores for patient selection and as a primary endpoint in clinical trials.

We introduce PAST-TIDE, our stance detection system addressing both subtasks of the StanceNakba Shared Task at NakbaNLP@LREC-COLING 2026. The main idea is statement tuning. We redefine stance as cloze-style masked language modeling (MLM), letting a verbalizer map label words to stance categories through the pre-trained MLM head rather than appending a randomly initialized classification head. We complement this with prototypical contrastive learning, which uses learnable class prototypes for batch-size independent contrastive training, and topic-conditional layer normalization for cross-topic Arabic stance detection. PAST-TIDE achieves macro-F1 scores of 0.75 for Subtask A and 0.74 for Subtask B on the official leaderboard, indicating that minimal architectural additions to a pre-trained model can remain competitive in low-resource settings.

Standard embodied evaluations do not independently score whether an agent correctly commits to task completion at episode closure, a capacity we call terminal commitment. Behaviorally distinct failures--never completing the task, completing it but failing to stop, and reporting success without sufficient evidence--collapse into the same benchmark failure. We introduce VIGIL, an evaluation framework that makes terminal commitment independently measurable. Under VIGIL's default protocol, agents observe only egocentric RGB, receive no action-success signals, and must end each episode with a semantic report checked deterministically against hidden world state. This yields two separate scores: world-state completion (W) and benchmark success (B), where B additionally requires a correct terminal report. This decoupling makes four outcome categories distinguishable: missed execution, post-attainment drift, unsupported commitment, and verified success. Across 20 models on 1,000 frozen episodes, systems with comparable W differ by up to 19.7 pp in B: one model converts achieved states into correct reports, while another with near-identical execution drifts past the goal without closing. An action-feedback intervention further tests the separation: execution-oriented signals improve W broadly, yet commitment failures persist in models that do not already ground terminal reports in the achieved state. VIGIL provides a protocol that makes terminal commitment independently visible and scorable.

In this work, we present two systems -- Named Entity Resolution (NER) and Natural Language Inference (NLI) -- for detecting legal violations within unstructured textual data and for associating these violations with potentially affected individuals, respectively. Both these systems are lightweight DeBERTa based encoders that outperform the LLM baselines. The proposed NER system achieved an F1 score of 60.01\% on Subtask A of the LegalLens challenge, which focuses on identifying violations. The proposed NLI system achieved an F1 score of 84.73\% on Subtask B of the LegalLens challenge, which focuses on resolving these violations by matching them with pre-existing legal complaints of class action cases. Our NER system ranked sixth and NLI system ranked fifth on the LegalLens leaderboard. We release the trained models and inference scripts.

Analyzing memes on the internet has emerged as a crucial endeavor due to the impact this multi-modal form of content wields in shaping online discourse. Memes have become a powerful tool for expressing emotions and sentiments, possibly even spreading hate and misinformation, through humor and sarcasm. In this paper, we present the overview of the Memotion 3 shared task, as part of the DeFactify 2 workshop at AAAI-23. The task released an annotated dataset of Hindi-English code-mixed memes based on their Sentiment (Task A), Emotion (Task B), and Emotion intensity (Task C). Each of these is defined as an individual task and the participants are ranked separately for each task. Over 50 teams registered for the shared task and 5 made final submissions to the test set of the Memotion 3 dataset. CLIP, BERT modifications, ViT etc. were the most popular models among the participants along with approaches such as Student-Teacher model, Fusion, and Ensembling. The best final F1 score for Task A is 34.41, Task B is 79.77 and Task C is 59.82.

In recent times, the detection of hate-speech, offensive, or abusive language in online media has become an important topic in NLP research due to the exponential growth of social media and the propagation of such messages, as well as their impact. Misogyny detection, even though it plays an important part in hate-speech detection, has not received the same attention. In this paper, we describe our classification systems submitted to the SemEval-2022 Task 5: MAMI - Multimedia Automatic Misogyny Identification. The shared task aimed to identify misogynous content in a multi-modal setting by analysing meme images together with their textual captions. To this end, we propose two models based on the pre-trained UNITER model, one enhanced with an image sentiment classifier, whereas the second leverages a Vocabulary Graph Convolutional Network (VGCN). Additionally, we explore an ensemble using the aforementioned models. Our best model reaches an F1-score of 71.4% in Sub-task A and 67.3% for Sub-task B positioning our team in the upper third of the leaderboard. We release the code and experiments for our models on GitHub

Machine learning is increasingly used to guide branch-and-cut (B&C) for mixed-integer linear programming by learning score-based policies for selecting branching variables and cutting planes. Many approaches train on local signals from lookahead heuristics such as strong branching, and linear programming (LP) bound improvement for cut selection. Training and evaluation of the learned models often focus on local score accuracy. We show that such local score-based methods can lead to search trees exponentially larger than optimal tree sizes, by identifying two sources of this gap. The first is that these widely used expert signals can be misaligned with overall tree size. LP bound improvement can select a root cut set that yields an exponentially larger strong branching tree than selecting cuts by a simple proxy score, and strong branching itself can be exponentially suboptimal (Dey et al., 2024). The second is that small discrepancies can be amplified by the branch-and-bound recursion. An arbitrarily small perturbation of the right-hand sides in a root cut set can change the minimum tree size from a single node to exponentially many. For branching, arbitrarily small score discrepancies, and differences only in tie-breaking, can produce trees of exponentially different sizes, and even a small number of decision differences along a trajectory can incur exponential growth. These results show that branch-and-cut policies trained and learned using local expert scores do not guarantee small trees, thus motivating the study of data-driven methods that produce policies better aligned with tree size rather than only accuracy on expert scores.

The parameter-efficient adaptation of the image-text pretraining model CLIP for video-text retrieval is a prominent area of research. While CLIP is focused on image-level vision-language matching, video-text retrieval demands comprehensive understanding at the video level. Three key discrepancies emerge in the transfer from image-level to video-level: vision, language, and alignment. However, existing methods mainly focus on vision while neglecting language and alignment. In this paper, we propose Discrepancy Reduction in Vision, Language, and Alignment (DiscoVLA), which simultaneously mitigates all three discrepancies. Specifically, we introduce Image-Video Features Fusion to integrate image-level and video-level features, effectively tackling both vision and language discrepancies. Additionally, we generate pseudo image captions to learn fine-grained image-level alignment. To mitigate alignment discrepancies, we propose Image-to-Video Alignment Distillation, which leverages image-level alignment knowledge to enhance video-level alignment. Extensive experiments demonstrate the superiority of our DiscoVLA. In particular, on MSRVTT with CLIP (ViT-B/16), DiscoVLA outperforms previous methods by 1.5% in R@1, reaching a final score of 50.5% R@1. The code is available at https://github.com/LunarShen/DsicoVLA.

Symbolic music datasets with matched scores and performances are essential for many music information retrieval (MIR) tasks. Yet, existing resources often cover a narrow range of composers, lack performance variety, omit note-level alignments, or use inconsistent naming formats. This work presents PianoCoRe, a large-scale piano MIDI dataset that unifies and refines major open-source piano corpora. The dataset contains 250,046 performances of 5,625 pieces written by 483 composers, totaling 21,763 h of performed music. PianoCoRe is released in tiered subsets to support different applications: from large-scale analysis and pre-training (PianoCoRe-C and deduplicated PianoCoRe-B) to expressive performance modeling with note-level score alignment (PianoCoRe-A/A*). The note-aligned subset, PianoCoRe-A, provides the largest open-source collection of 157,207 performances aligned to 1,591 scores to date. In addition to the dataset, the contributions are: (1) a MIDI quality classifier for detecting corrupted and score-like transcriptions and (2) RAScoP, an alignment refinement pipeline that cleans temporal alignment errors and interpolates missing notes. The analysis shows that the refinement reduces temporal noise and eliminates tempo outliers. Moreover, an expressive performance rendering model trained on PianoCoRe demonstrates improved robustness to unseen pieces compared to models trained on raw or smaller datasets. PianoCoRe provides a ready-to-use foundation for the next generation of expressive piano performance research.

LLM-as-a-judge has emerged as a cornerstone technique for evaluating large language models by leveraging LLM reasoning to score prompt-response pairs. Since LLM judgments are stochastic, practitioners commonly query each pair multiple times to estimate mean scores accurately. This raises a critical challenge: given a fixed computational budget B, how to optimally allocate queries across K prompt-response pairs to minimize estimation error? % We present a principled variance-adaptive approach leveraging multi-armed bandit theory and concentration inequalities. Our method dynamically allocates queries based on estimated score variances, concentrating resources where uncertainty is highest. Further, our algorithm is shown to achieve a worst-case score-estimation error of Oleft(frac{sum_{i=1^K σ_i^2}{B}}right), σ_i^2 being the unknown score variance for pair i in [K] with near-optimal budget allocation. % Experiments on Summarize-From-Feedback and HelpSteer2 demonstrate that our method significantly outperforms uniform allocation, reducing worst-case estimation error while maintaining identical budgets. Our work establishes a theoretical foundation for efficient LLM evaluation with practical implications for AI safety, model alignment, and automated assessment at scale.

As Large Language Models (LLMs) are increasingly deployed as autonomous agents in complex and long horizon settings, it is critical to evaluate their ability to sabotage users by pursuing hidden objectives. We study the ability of frontier LLMs to evade monitoring and achieve harmful hidden goals while completing a wide array of realistic tasks. We evaluate a broad range of frontier LLMs using SHADE (Subtle Harmful Agent Detection & Evaluation)-Arena, the first highly diverse agent evaluation dataset for sabotage and monitoring capabilities of LLM agents. SHADE-Arena consists of complex pairs of benign main tasks and harmful side objectives in complicated environments. Agents are evaluated on their ability to complete the side task without appearing suspicious to an LLM monitor. When measuring agent ability to (a) complete the main task, (b) complete the side task, and (c) avoid detection, we find that the best performing frontier models score 27% (Claude 3.7 Sonnet) and 15% (Gemini 2.5 Pro) as sabotage agents when overseen by Claude 3.6 Sonnet. For current frontier models, success on the side task relies heavily on having access to a hidden scratchpad that is not visible to the monitor. We also use SHADE-Arena to measure models' monitoring abilities, with the top monitor (Gemini 2.5 Pro) achieving an AUC of 0.87 at distinguishing benign and malign transcripts. We find that for now, models still struggle at sabotage due to failures in long-context main task execution. However, our measurements already demonstrate the difficulty of monitoring for subtle sabotage attempts, which we expect to only increase in the face of more complex and longer-horizon tasks.

Self-supervised pretraining on large-scale satellite data has raised great interest in building Earth observation (EO) foundation models. However, many important resources beyond pure satellite imagery, such as land-cover-land-use products that provide free global semantic information, as well as vision foundation models that hold strong knowledge of the natural world, tend to be overlooked. In this work, we show these free additional resources not only help resolve common contrastive learning bottlenecks, but also significantly boost the efficiency and effectiveness of EO pretraining. Specifically, we first propose soft contrastive learning that optimizes cross-scene soft similarity based on land-cover-generated multi-label supervision, naturally solving the issue of multiple positive samples and too strict positive matching in complex scenes. Second, we explore cross-domain continual pretraining for both multispectral and SAR imagery, building efficient EO foundation models from strongest vision models such as DINOv2. Integrating simple weight-initialization and Siamese masking strategies into our soft contrastive learning framework, we demonstrate impressive continual pretraining performance even when the input channels and modalities are not aligned. Without prohibitive training, we produce multispectral and SAR foundation models that achieve significantly better results in 9 out of 10 downstream tasks than most existing SOTA models. For example, our ResNet50/ViT-S achieve 84.8/85.0 linear probing mAP scores on BigEarthNet-10\% which are better than most existing ViT-L models; under the same setting, our ViT-B sets a new record of 86.8 in multispectral, and 82.5 in SAR, the latter even better than many multispectral models. Dataset and models are available at https://github.com/zhu-xlab/softcon.

Automatic dialogue coherence evaluation has attracted increasing attention and is crucial for developing promising dialogue systems. However, existing metrics have two major limitations: (a) they are mostly trained in a simplified two-level setting (coherent vs. incoherent), while humans give Likert-type multi-level coherence scores, dubbed as "quantifiable"; (b) their predicted coherence scores cannot align with the actual human rating standards due to the absence of human guidance during training. To address these limitations, we propose Quantifiable Dialogue Coherence Evaluation (QuantiDCE), a novel framework aiming to train a quantifiable dialogue coherence metric that can reflect the actual human rating standards. Specifically, QuantiDCE includes two training stages, Multi-Level Ranking (MLR) pre-training and Knowledge Distillation (KD) fine-tuning. During MLR pre-training, a new MLR loss is proposed for enabling the model to learn the coarse judgement of coherence degrees. Then, during KD fine-tuning, the pretrained model is further finetuned to learn the actual human rating standards with only very few human-annotated data. To advocate the generalizability even with limited fine-tuning data, a novel KD regularization is introduced to retain the knowledge learned at the pre-training stage. Experimental results show that the model trained by QuantiDCE presents stronger correlations with human judgements than the other state-of-the-art metrics.

We introduce SWE-PRBench, a benchmark of 350 pull requests with human-annotated ground truth for evaluating AI code review quality. Evaluated against an LLM-as-judge framework validated at kappa=0.75, 8 frontier models detect only 15-31% of human-flagged issues on the diff-only configuration, demonstrating that AI code review remains far below human expert performance despite strong results on code generation benchmarks. Pull requests are drawn from active open-source repositories, filtered from 700 candidates using a Repository Quality Score, and evaluated under three frozen context configurations: diff only (config_A), diff with file content (config_B), and full context (config_C), enabling systematic ablation of context provision strategies. All 8 models degrade monotonically from config_A to config_C, even when context is provided via structured semantic layers including AST-extracted function context and import graph resolution. The dominant mechanism is a collapse of Type2_Contextual issue detection at config_B, consistent with attention dilution in long contexts: a structured 2,000-token diff-with-summary prompt outperforms a 2,500-token full-context prompt enriched with execution context, behaviour mapping, and test signatures across all 8 models. The top four models are statistically indistinguishable (mean score 0.147-0.153) while a clear tier gap separates them from the remaining four (mean score <= 0.113). Dataset, contexts, annotations, and evaluation harness are released publicly.

We present GaRAGe, a large RAG benchmark with human-curated long-form answers and annotations of each grounding passage, allowing a fine-grained evaluation of whether LLMs can identify relevant grounding when generating RAG answers. Our benchmark contains 2366 questions of diverse complexity, dynamism, and topics, and includes over 35K annotated passages retrieved from both private document sets and the Web, to reflect real-world RAG use cases. This makes it an ideal test bed to evaluate an LLM's ability to identify only the relevant information necessary to compose a response, or provide a deflective response when there is insufficient information. Evaluations of multiple state-of-the-art LLMs on GaRAGe show that the models tend to over-summarise rather than (a) ground their answers strictly on the annotated relevant passages (reaching at most a Relevance-Aware Factuality Score of 60%), or (b) deflect when no relevant grounding is available (reaching at most 31% true positive rate in deflections). The F1 in attribution to relevant sources is at most 58.9%, and we show that performance is particularly reduced when answering time-sensitive questions and when having to draw knowledge from sparser private grounding sources.

Aligning large language models (LLMs) with human values and intents critically involves the use of human or AI feedback. While dense feedback annotations are expensive to acquire and integrate, sparse feedback presents a structural design choice between ratings (e.g., score Response A on a scale of 1-7) and rankings (e.g., is Response A better than Response B?). In this work, we analyze the effect of this design choice for the alignment and evaluation of LLMs. We uncover an inconsistency problem wherein the preferences inferred from ratings and rankings significantly disagree 60% for both human and AI annotators. Our subsequent analysis identifies various facets of annotator biases that explain this phenomena, such as human annotators would rate denser responses higher while preferring accuracy during pairwise judgments. To our surprise, we also observe that the choice of feedback protocol also has a significant effect on the evaluation of aligned LLMs. In particular, we find that LLMs that leverage rankings data for alignment (say model X) are preferred over those that leverage ratings data (say model Y), with a rank-based evaluation protocol (is X/Y's response better than reference response?) but not with a rating-based evaluation protocol (score Rank X/Y's response on a scale of 1-7). Our findings thus shed light on critical gaps in methods for evaluating the real-world utility of language models and their strong dependence on the feedback protocol used for alignment. Our code and data are available at https://github.com/Hritikbansal/sparse_feedback.

Creating programs to correctly manipulate data is a difficult task, as the underlying programming languages and APIs can be challenging to learn for many users who are not skilled programmers. Large language models (LLMs) demonstrate remarkable potential for generating code from natural language, but in the data manipulation domain, apart from the natural language (NL) description of the intended task, we also have the dataset on which the task is to be performed, or the "data context". Existing approaches have utilized data context in a limited way by simply adding relevant information from the input data into the prompts sent to the LLM. In this work, we utilize the available input data to execute the candidate programs generated by the LLMs and gather their outputs. We introduce semantic reranking, a technique to rerank the programs generated by LLMs based on three signals coming the program outputs: (a) semantic filtering and well-formedness based score tuning: do programs even generate well-formed outputs, (b) semantic interleaving: how do the outputs from different candidates compare to each other, and (c) output-based score tuning: how do the outputs compare to outputs predicted for the same task. We provide theoretical justification for semantic interleaving. We also introduce temperature mixing, where we combine samples generated by LLMs using both high and low temperatures. We extensively evaluate our approach in three domains, namely databases (SQL), data science (Pandas) and business intelligence (Excel's Power Query M) on a variety of new and existing benchmarks. We observe substantial gains across domains, with improvements of up to 45% in top-1 accuracy and 34% in top-3 accuracy.

Recent state-of-the-art natural language understanding models, such as BERT and XLNet, score a pair of sentences (A and B) using multiple cross-attention operations - a process in which each word in sentence A attends to all words in sentence B and vice versa. As a result, computing the similarity between a query sentence and a set of candidate sentences, requires the propagation of all query-candidate sentence-pairs throughout a stack of cross-attention layers. This exhaustive process becomes computationally prohibitive when the number of candidate sentences is large. In contrast, sentence embedding techniques learn a sentence-to-vector mapping and compute the similarity between the sentence vectors via simple elementary operations. In this paper, we introduce Distilled Sentence Embedding (DSE) - a model that is based on knowledge distillation from cross-attentive models, focusing on sentence-pair tasks. The outline of DSE is as follows: Given a cross-attentive teacher model (e.g. a fine-tuned BERT), we train a sentence embedding based student model to reconstruct the sentence-pair scores obtained by the teacher model. We empirically demonstrate the effectiveness of DSE on five GLUE sentence-pair tasks. DSE significantly outperforms several ELMO variants and other sentence embedding methods, while accelerating computation of the query-candidate sentence-pairs similarities by several orders of magnitude, with an average relative degradation of 4.6% compared to BERT. Furthermore, we show that DSE produces sentence embeddings that reach state-of-the-art performance on universal sentence representation benchmarks. Our code is made publicly available at https://github.com/microsoft/Distilled-Sentence-Embedding.

Recent multimodal LLMs have shown promise in chart-based visual question answering, but their performance declines sharply on unannotated charts, those requiring precise visual interpretation rather than relying on textual shortcuts. To address this, we introduce ChartAgent, a novel agentic framework that explicitly performs visual reasoning directly within the chart's spatial domain. Unlike textual chain-of-thought reasoning, ChartAgent iteratively decomposes queries into visual subtasks and actively manipulates and interacts with chart images through specialized actions such as drawing annotations, cropping regions (e.g., segmenting pie slices, isolating bars), and localizing axes, using a library of chart-specific vision tools to fulfill each subtask. This iterative reasoning process closely mirrors human cognitive strategies for chart comprehension. ChartAgent achieves state-of-the-art accuracy on the ChartBench and ChartX benchmarks, surpassing prior methods by up to 16.07% absolute gain overall and 17.31% on unannotated, numerically intensive queries. Furthermore, our analyses show that ChartAgent is (a) effective across diverse chart types, (b) achieve the highest scores across varying visual and reasoning complexity levels, and (c) serves as a plug-and-play framework that boosts performance across diverse underlying LLMs. Our work is among the first to demonstrate visually grounded reasoning for chart understanding using tool-augmented multimodal agents.

Compositional generalization is crucial for artificial intelligence agents to solve complex vision-language reasoning tasks. Neuro-symbolic approaches have demonstrated promise in capturing compositional structures, but they face critical challenges: (a) reliance on predefined predicates for symbolic representations that limit adaptability, (b) difficulty in extracting predicates from raw data, and (c) using non-differentiable operations for combining primitive concepts. To address these issues, we propose NeSyCoCo, a neuro-symbolic framework that leverages large language models (LLMs) to generate symbolic representations and map them to differentiable neural computations. NeSyCoCo introduces three innovations: (a) augmenting natural language inputs with dependency structures to enhance the alignment with symbolic representations, (b) employing distributed word representations to link diverse, linguistically motivated logical predicates to neural modules, and (c) using the soft composition of normalized predicate scores to align symbolic and differentiable reasoning. Our framework achieves state-of-the-art results on the ReaSCAN and CLEVR-CoGenT compositional generalization benchmarks and demonstrates robust performance with novel concepts in the CLEVR-SYN benchmark.

Reasoning is a cognitive process of using evidence to reach a sound conclusion. The reasoning capability is essential for large language models (LLMs) to serve as the brain of the artificial general intelligence agent. Recent studies reveal that fine-tuning LLMs on data with the chain of thought (COT) reasoning process can significantly enhance their reasoning capabilities. However, we find that the fine-tuned LLMs suffer from an Assessment Misalignment problem, i.e., they frequently assign higher scores to subpar COTs, leading to potential limitations in their reasoning abilities. To address this problem, we introduce an Alignment Fine-Tuning (AFT) paradigm, which involves three steps: 1) fine-tuning LLMs with COT training data; 2) generating multiple COT responses for each question, and categorizing them into positive and negative ones based on whether they achieve the correct answer; 3) calibrating the scores of positive and negative responses given by LLMs with a novel constraint alignment loss. Specifically, the constraint alignment loss has two objectives: a) Alignment, which guarantees that positive scores surpass negative scores to encourage answers with high-quality COTs; b) Constraint, which keeps the negative scores confined to a reasonable range to prevent the model degradation. Beyond just the binary positive and negative feedback, the constraint alignment loss can be seamlessly adapted to the ranking situations when ranking feedback is accessible. Furthermore, we also delve deeply into recent ranking-based alignment methods, such as DPO, RRHF, and PRO, and discover that the constraint, which has been overlooked by these approaches, is also crucial for their performance. Extensive experiments on four reasoning benchmarks with both binary and ranking feedback demonstrate the effectiveness of AFT.

Multi-hop retrieval is not a single-step relevance problem: later-hop evidence should be ranked by its utility conditioned on retrieved bridge evidence, not by similarity to the original query alone. We present BridgeRAG, a training-free, graph-free retrieval method for retrieval-augmented generation (RAG) over multi-hop questions that operationalizes this view with a tripartite scorer s(q,b,c) over (question, bridge, candidate). BridgeRAG separates coverage from scoring: dual-entity ANN expansion broadens the second-hop candidate pool, while a bridge-conditioned LLM judge identifies the active reasoning chain among competing candidates without any offline graph or proposition index. Across four controlled experiments we show that this conditioning signal is (i) selective: +2.55pp on parallel-chain queries (p<0.001) vs. ~0 on single-chain subtypes; (ii) irreplaceable: substituting the retrieved passage with generated SVO query text reduces R@5 by 2.1pp, performing worse than even the lowest-SVO-similarity pool passage; (iii) predictable: cos(b,g2) correlates with per-query gain (Spearman rho=0.104, p<0.001); and (iv) mechanistically precise: bridge conditioning causes productive re-rankings (18.7% flip-win rate on parallel-chain vs. 0.6% on single-chain), not merely more churn. Combined with lightweight coverage expansion and percentile-rank score fusion, BridgeRAG achieves the best published training-free R@5 under matched benchmark evaluation on all three standard MHQA benchmarks without a graph database or any training: 0.8146 on MuSiQue (+3.1pp vs. PropRAG, +6.8pp vs. HippoRAG2), 0.9527 on 2WikiMultiHopQA (+1.2pp vs. PropRAG), and 0.9875 on HotpotQA (+1.35pp vs. PropRAG).

Mixture-of-Experts large language models (MoE-LLMs) marks a significant step forward of language models, however, they encounter two critical challenges in practice: 1) expert parameters lead to considerable memory consumption and loading latency; and 2) the current activated experts are redundant, as many tokens may only require a single expert. Motivated by these issues, we investigate the MoE-LLMs and make two key observations: a) different experts exhibit varying behaviors on activation reconstruction error, routing scores, and activated frequencies, highlighting their differing importance, and b) not all tokens are equally important -- only a small subset is critical. Building on these insights, we propose MC-MoE, a training-free Mixture-Compressor for MoE-LLMs, which leverages the significance of both experts and tokens to achieve an extreme compression. First, to mitigate storage and loading overheads, we introduce Pre-Loading Mixed-Precision Quantization, which formulates the adaptive bit-width allocation as a Linear Programming problem, where the objective function balances multi-factors reflecting the importance of each expert. Additionally, we develop Online Dynamic Pruning, which identifies important tokens to retain and dynamically select activated experts for other tokens during inference to optimize efficiency while maintaining performance. Our MC-MoE integrates static quantization and dynamic pruning to collaboratively achieve extreme compression for MoE-LLMs with less accuracy loss, ensuring an optimal trade-off between performance and efficiency. Extensive experiments confirm the effectiveness of our approach. For instance, at 2.54 bits, MC-MoE compresses 76.6% of the model, with only a 3.8% average accuracy loss. During dynamic inference, we further reduce activated parameters by 15%, with a performance drop of less than 0.6%.

In disentangling the heterogeneity observed in psychopathology, personality of the patients is considered crucial. While it has been demonstrated that personality traits are reflected in the language used by a patient, we hypothesize that this enables automatic inference of the personality type directly from speech utterances, potentially more accurately than through a traditional questionnaire-based approach explicitly designed for personality classification. To validate this hypothesis, we adopt natural language processing (NLP) and standard machine learning tools for classification. We test this on a dataset of recorded clinical diagnostic interviews (CDI) on a sample of 79 patients diagnosed with major depressive disorder (MDD) -- a condition for which differentiated treatment based on personality styles has been advocated -- and classified into anaclitic and introjective personality styles. We start by analyzing the interviews to see which linguistic features are associated with each style, in order to gain a better understanding of the styles. Then, we develop automatic classifiers based on (a) standardized questionnaire responses; (b) basic text features, i.e., TF-IDF scores of words and word sequences; (c) more advanced text features, using LIWC (linguistic inquiry and word count) and context-aware features using BERT (bidirectional encoder representations from transformers); (d) audio features. We find that automated classification with language-derived features (i.e., based on LIWC) significantly outperforms questionnaire-based classification models. Furthermore, the best performance is achieved by combining LIWC with the questionnaire features. This suggests that more work should be put into developing linguistically based automated techniques for characterizing personality, however questionnaires still to some extent complement such methods.

We present LogiGAN, an unsupervised adversarial pre-training framework for improving logical reasoning abilities of language models. Upon automatic identifying logical reasoning phenomena in massive text corpus via detection heuristics, we train language models to predict the masked-out logical statements. Inspired by the facilitation effect of reflective thinking in human learning, we analogically simulate the learning-thinking process with an adversarial Generator-Verifier architecture to assist logic learning. LogiGAN implements a novel sequential GAN approach that (a) circumvents the non-differentiable challenge of the sequential GAN by leveraging the Generator as a sentence-level generative likelihood scorer with a learning objective of reaching scoring consensus with the Verifier; (b) is computationally feasible for large-scale pre-training with arbitrary target length. Both base and large size language models pre-trained with LogiGAN demonstrate obvious performance improvement on 12 datasets requiring general reasoning abilities, revealing the fundamental role of logic in broad reasoning, as well as the effectiveness of LogiGAN. Ablation studies on LogiGAN components reveal the relative orthogonality between linguistic and logic abilities and suggest that reflective thinking's facilitation effect might also generalize to machine learning.

Weakly supervised object localization (WSOL) relaxes the requirement of dense annotations for object localization by using image-level classification masks to supervise its learning process. However, current WSOL methods suffer from excessive activation of background locations and need post-processing to obtain the localization mask. This paper attributes these issues to the unawareness of background cues, and propose the background-aware classification activation map (B-CAM) to simultaneously learn localization scores of both object and background with only image-level labels. In our B-CAM, two image-level features, aggregated by pixel-level features of potential background and object locations, are used to purify the object feature from the object-related background and to represent the feature of the pure-background sample, respectively. Then based on these two features, both the object classifier and the background classifier are learned to determine the binary object localization mask. Our B-CAM can be trained in end-to-end manner based on a proposed stagger classification loss, which not only improves the objects localization but also suppresses the background activation. Experiments show that our B-CAM outperforms one-stage WSOL methods on the CUB-200, OpenImages and VOC2012 datasets.

Reinforcement Learning with Verifiable Rewards (RLVR) significantly enhances the reasoning capability of Large Language Models (LLMs). Current RLVR approaches typically conduct training across all generated tokens, but neglect to explore which tokens (e.g., prefix tokens) actually contribute to reasoning. This uniform training strategy spends substantial effort on optimizing low-return tokens, which in turn impedes the potential improvement from high-return tokens and reduces overall training effectiveness. To address this issue, we propose a novel RLVR approach called Progressive Prefix-token Policy Optimization (PPPO), which highlights the significance of the prefix segment of generated outputs. Specifically, inspired by the well-established human thinking theory of Path Dependence, where early-stage thoughts substantially constrain subsequent thinking trajectory, we identify an analogous phenomenon in LLM reasoning termed Beginning Lock-in Effect (BLE). PPPO leverages this finding by focusing its optimization objective on the prefix reasoning process of LLMs. This targeted optimization strategy can positively influence subsequent reasoning processes, and ultimately improve final results. To improve the learning effectiveness of LLMs on how to start reasoning with high quality, PPPO introduces two training strategies: (a) Progressive Prefix Retention, which shapes a progressive learning process by increasing the proportion of retained prefix tokens during training; (b) Continuation Accumulated Reward, which mitigates reward bias by sampling multiple continuations for one prefix token sequence, and accumulating their scores as the reward signal. Extensive experimental results on various reasoning tasks demonstrate that our proposed PPPO outperforms representative RLVR methods, with the accuracy improvements of 18.02% on only 26.17% training tokens.

In recent years, advancements in Natural Language Processing (NLP) techniques have revolutionized the field of accessibility and exclusivity of testing, particularly for visually impaired students (VIS). CBT has shown in years back its relevance in terms of administering exams electronically, making the test process easier, providing quicker and more accurate results, and offering greater flexibility and accessibility for candidates. Yet, its relevance was not felt by the visually impaired students as they cannot access printed documents. Hence, in this paper, we present an NLP-driven Computer-Based Test guide for visually impaired students. It employs a speech technology pre-trained methods to provide real-time assistance and support to visually impaired students. The system utilizes NLP technologies to convert the text-based questions and the associated options in a machine-readable format. Subsequently, the speech technology pre-trained model processes the converted text enabling the VIS to comprehend and analyze the content. Furthermore, we validated that this pre-trained model is not perverse by testing for accuracy using sample audio datasets labels (A, B, C, D, E, F, G) to compare with the voice recordings obtained from 20 VIS which is been predicted by the system to attain values for precision, recall, and F1-scores. These metrics are used to assess the performance of the pre-trained model and have indicated that it is proficient enough to give its better performance to the evaluated system. The methodology adopted for this system is Object Oriented Analysis and Design Methodology (OOADM) where Objects are discussed and built by modeling real-world instances.

Prompt Tuning is emerging as a scalable and cost-effective method to fine-tune Pretrained Language Models (PLMs), which are often referred to as Large Language Models (LLMs). This study benchmarks the performance and computational efficiency of Prompt Tuning and baselines for multi-label text classification. This is applied to the challenging task of classifying companies into an investment firm's proprietary industry taxonomy, supporting their thematic investment strategy. Text-to-text classification is frequently reported to outperform task-specific classification heads, but has several limitations when applied to a multi-label classification problem where each label consists of multiple tokens: (a) Generated labels may not match any label in the label taxonomy; (b) The fine-tuning process lacks permutation invariance and is sensitive to the order of the provided labels; (c) The model provides binary decisions rather than appropriate confidence scores. Limitation (a) is addressed by applying constrained decoding using Trie Search, which slightly improves classification performance. All limitations (a), (b), and (c) are addressed by replacing the PLM's language head with a classification head, which is referred to as Prompt Tuned Embedding Classification (PTEC). This improves performance significantly, while also reducing computational costs during inference. In our industrial application, the training data is skewed towards well-known companies. We confirm that the model's performance is consistent across both well-known and less-known companies. Our overall results indicate the continuing need to adapt state-of-the-art methods to domain-specific tasks, even in the era of PLMs with strong generalization abilities. We release our codebase and a benchmarking dataset at https://github.com/EQTPartners/PTEC.

Agentic large language model (LLM) systems rely on external memory for long-horizon state and concurrent multi-agent execution, but centralized indexes and heuristic partitions become bottlenecks as memory volume and parallel access grow. We present ShardMemo, a budgeted tiered memory service with Tier A per-agent working state, Tier B sharded evidence with shard-local approximate nearest neighbor (ANN) indexes, and Tier C, a versioned skill library. Tier B enforces scope-before-routing: structured eligibility constraints mask ineligible shards before routing or ANN search. We cast shard probing as masked mixture-of-experts (MoE) routing over eligible shards, probing up to B_{probe} shards via Top-B_{probe} or adaptive Top-P, and use cost-aware gating over profile/observation/session shard families; the router is trained from evidence-to-shard supervision. On LoCoMo, ShardMemo improves over the strongest baseline (GAM) by +5.11 to +6.82 F1 across question categories. Under a fixed-budget routing setting (B_{probe}=3), ShardMemo improves over cosine-to-prototype shard routing by +6.87 F1 while reducing retrieval work (VecScan 521->414, -20.5%) and p95 latency (95->76 ms). On long-context HotpotQA, ShardMemo achieves 63.41/61.88/57.95 F1 at 56K/224K/448K tokens. On ToolBench, Tier C reaches 0.97 Precision@3 and 1.94 StepRed (+10.2% and +7.2% over embedding-similarity retrieval).

Deep learning-based diagnostic models often suffer performance drops due to distribution shifts between training (source) and test (target) domains. Collecting and labeling sufficient target domain data for model retraining represents an optimal solution, yet is limited by time and scarce resources. Active learning (AL) offers an efficient approach to reduce annotation costs while maintaining performance, but struggles to handle the challenge posed by distribution variations across different datasets. In this study, we propose a novel unsupervised Active learning framework for Domain Adaptation, named ADAptation, which efficiently selects informative samples from multi-domain data pools under limited annotation budget. As a fundamental step, our method first utilizes the distribution homogenization capabilities of diffusion models to bridge cross-dataset gaps by translating target images into source-domain style. We then introduce two key innovations: (a) a hypersphere-constrained contrastive learning network for compact feature clustering, and (b) a dual-scoring mechanism that quantifies and balances sample uncertainty and representativeness. Extensive experiments on four breast ultrasound datasets (three public and one in-house/multi-center) across five common deep classifiers demonstrate that our method surpasses existing strong AL-based competitors, validating its effectiveness and generalization for clinical domain adaptation. The code is available at the anonymized link: https://github.com/miccai25-966/ADAptation.

Generative models trained using self-supervision of tokenized electronic health record (EHR) timelines show promise for clinical outcome prediction. This is typically done using Monte Carlo simulation for future patient trajectories. However, existing approaches suffer from three key limitations: sparse estimate distributions that poorly differentiate patient risk levels, extreme computational costs, and high sampling variance. We propose two new estimators: the Sum of Conditional Outcome Probability Estimator (SCOPE) and Risk Estimation from Anticipated Conditional Hazards (REACH), that leverage next-token probability distributions discarded by standard Monte Carlo. We prove both estimators are unbiased and that REACH guarantees variance reduction over Monte Carlo sampling for any model and outcome. Empirically, on hospital mortality prediction in MIMIC-IV using the ETHOS-ARES framework, SCOPE and REACH match 100-sample Monte Carlo performance using only 10-11 samples (95% CI: [9,11]), representing a ~10x reduction in inference cost without degrading calibration. For ICU admission prediction, efficiency gains are more modest (~1.2x), which we attribute to the outcome's lower "spontaneity," a property we characterize theoretically and empirically. These methods substantially improve the feasibility of deploying generative EHR models in resource-constrained clinical settings.

Some speech recognition tasks, such as automatic speech recognition (ASR), are approaching or have reached human performance in many reported metrics. Yet, they continue to struggle in complex, real-world, situations, such as with distanced speech. Previous challenges have released datasets to address the issue of distanced ASR, however, the focus remains primarily on distance, specifically relying on multi-microphone array systems. Here we present the B(asic) E(motion) R(andom phrase) S(hou)t(s) (BERSt) dataset. The dataset contains almost 4 hours of English speech from 98 actors with varying regional and non-native accents. The data was collected on smartphones in the actors homes and therefore includes at least 98 different acoustic environments. The data also includes 7 different emotion prompts and both shouted and spoken utterances. The smartphones were places in 19 different positions, including obstructions and being in a different room than the actor. This data is publicly available for use and can be used to evaluate a variety of speech recognition tasks, including: ASR, shout detection, and speech emotion recognition (SER). We provide initial benchmarks for ASR and SER tasks, and find that ASR degrades both with an increase in distance and shout level and shows varied performance depending on the intended emotion. Our results show that the BERSt dataset is challenging for both ASR and SER tasks and continued work is needed to improve the robustness of such systems for more accurate real-world use.

Semantic textual similartiy (STS) and information retrieval tasks (IR) tasks have been the two major avenues to record the progress of embedding models in the past few years. Under the emerging Retrieval-augmented Generation (RAG) paradigm, we envision the need to evaluate next-level language understanding abilities of embedding models, and take a conscious look at the reasoning abilities stored in them. Addressing this, we pose the question: Can retrievers solve reasoning problems? By transforming reasoning tasks into retrieval tasks, we find that without specifically trained for reasoning-level language understanding, current state-of-the-art retriever models may still be far from being competent for playing the role of assisting LLMs, especially in reasoning-intensive tasks. Moreover, albeit trained to be aware of instructions, instruction-aware IR models are often better off without instructions in inference time for reasoning tasks, posing an overlooked retriever-LLM behavioral gap for the research community to align. However, recent decoder-based embedding models show great promise in narrowing the gap, highlighting the pathway for embedding models to achieve reasoning-level language understanding. We also show that, although current off-the-shelf re-ranker models fail on these tasks, injecting reasoning abilities into them through fine-tuning still appears easier than doing so to bi-encoders, and we are able to achieve state-of-the-art performance across all tasks by fine-tuning a reranking model. We release Reasoning as Retrieval Benchmark (RAR-b), a holistic suite of tasks and settings to evaluate the reasoning abilities stored in retriever models. RAR-b is available at https://github.com/gowitheflow-1998/RAR-b.

Recent advancements in large language models (LLMs) have propelled Artificial Intelligence (AI) to new heights, enabling breakthroughs in various tasks such as writing assistance, code generation, and machine translation. A significant distinction of advanced LLMs, such as ChatGPT, is their demonstrated ability to "reason." However, evaluating the reasoning ability of LLMs remains a challenge as most existing evaluations focus on their accuracy on the downstream tasks rather than directly assessing their reasoning processes. Efforts have been made to develop benchmarks and metrics to assess reasoning in LLMs, but they suffer from data leakage or limited scope. In this paper, we introduce LogicAsker, an automatic approach that comprehensively evaluates and improves the logical reasoning abilities of LLMs under a set of atomic reasoning skills based on propositional and predicate logic. The results provide insights into LLMs' reasoning abilities and reveal the logical rules the LLMs did not learn well. We evaluate LogicAsker on six widely deployed LLMs, including GPT-3, ChatGPT, GPT-4, Bard, Vicuna, and Guanaco. The results show that test cases from LogicAsker can find logical reasoning failures in different LLMs with a rate of 25\% - 94\%. In addition, the test cases of LogicAsker can be further used to design demonstration examples for in-context learning, which effectively improves the logical reasoning ability of LLMs, e.g., 10\% for GPT-4. As far as we know, our work is the first to create prompts based on testing results to improve LLMs' formal reasoning ability effectively. All the code, data, and results will be released for reproduction and future research.

Reliable retrieval-augmented generation (RAG) systems depend fundamentally on the retriever's ability to find relevant information. We show that neural retrievers used in RAG systems have blind spots, which we define as the failure to retrieve entities that are relevant to the query, but have low similarity to the query embedding. We investigate the training-induced biases that cause such blind spot entities to be mapped to inaccessible parts of the embedding space, resulting in low retrievability. Using a large-scale dataset constructed from Wikidata relations and first paragraphs of Wikipedia, and our proposed Retrieval Probability Score (RPS), we show that blind spot risk in standard retrievers (e.g., CONTRIEVER, REASONIR) can be predicted pre-index from entity embedding geometry, avoiding expensive retrieval evaluations. To address these blind spots, we introduce ARGUS, a pipeline that enables the retrievability of high-risk (low-RPS) entities through targeted document augmentation from a knowledge base (KB), first paragraphs of Wikipedia, in our case. Extensive experiments on BRIGHT, IMPLIRET, and RAR-B show that ARGUS achieves consistent improvements across all evaluated retrievers (averaging +3.4 nDCG@5 and +4.5 nDCG@10 absolute points), with substantially larger gains in challenging subsets. These results establish that preemptively remedying blind spots is critical for building robust and trustworthy RAG systems (Code and Data).

Evolving borrower behaviors, shifting economic conditions, and changing regulatory landscapes continuously reshape the data distributions underlying modern credit-scoring systems. Conventional explainability techniques, such as SHAP, assume static data and fixed background distributions, making their explanations unstable and potentially unfair when concept drift occurs. This study addresses that challenge by developing adaptive explanation frameworks that recalibrate interpretability and fairness in dynamically evolving credit models. Using a multi-year credit dataset, we integrate predictive modeling via XGBoost with three adaptive SHAP variants: (A) per-slice explanation reweighting that adjusts for feature distribution shifts, (B) drift-aware SHAP rebaselining with sliding-window background samples, and (C) online surrogate calibration using incremental Ridge regression. Each method is benchmarked against static SHAP explanations using metrics of predictive performance (AUC, F1), directional and rank stability (cosine, Kendall tau), and fairness (demographic parity and recalibration). Results show that adaptive methods, particularly rebaselined and surrogate-based explanations, substantially improve temporal stability and reduce disparate impact across demographic groups without degrading predictive accuracy. Robustness tests, including counterfactual perturbations, background sensitivity analysis, and proxy-variable detection, confirm the resilience of adaptive explanations under real-world drift conditions. These findings establish adaptive explainability as a practical mechanism for sustaining transparency, accountability, and ethical reliability in data-driven credit systems, and more broadly, in any domain where decision models evolve with population change.

Skiplists are widely used for in-memory indexing in many key-value stores, such as RocksDB and LevelDB, due to their ease of implementation and simple concurrency control mechanisms. However, traditional skiplists suffer from poor cache locality, as they store only a single element per node, leaving performance on the table. Minimizing last-level cache misses is key to maximizing in-memory index performance, making high cache locality essential. In this paper, we present a practical concurrent B-skiplist that enhances cache locality and performance while preserving the simplicity of traditional skiplist structures and concurrency control schemes. Our key contributions include a top-down, single-pass insertion algorithm for B-skiplists and a corresponding simple and efficient top-down concurrency control scheme. On 128 threads, the proposed concurrent B-skiplist achieves between 2x-9x higher throughput compared to state-of-the-art concurrent skiplist implementations, including Facebook's concurrent skiplist from Folly and the Java ConcurrentSkipListMap. Furthermore, we find that the B-skiplist achieves competitive (0.9x-1.7x) throughput on point workloads compared to state-of-the-art cache-optimized tree-based indices (e.g., Masstree). For a more complete picture of the performance, we also measure the latency of skiplist and tree-based indices and find that the B-skiplist achieves between 3.5x-103x lower 99% latency compared to other concurrent skiplists and between 0.85x-64x lower 99% latency compared to tree-based indices on point workloads with inserts.

Coaching, which involves classroom observation and expert feedback, is a widespread and fundamental part of teacher training. However, the majority of teachers do not have access to consistent, high quality coaching due to limited resources and access to expertise. We explore whether generative AI could become a cost-effective complement to expert feedback by serving as an automated teacher coach. In doing so, we propose three teacher coaching tasks for generative AI: (A) scoring transcript segments based on classroom observation instruments, (B) identifying highlights and missed opportunities for good instructional strategies, and (C) providing actionable suggestions for eliciting more student reasoning. We recruit expert math teachers to evaluate the zero-shot performance of ChatGPT on each of these tasks for elementary math classroom transcripts. Our results reveal that ChatGPT generates responses that are relevant to improving instruction, but they are often not novel or insightful. For example, 82% of the model's suggestions point to places in the transcript where the teacher is already implementing that suggestion. Our work highlights the challenges of producing insightful, novel and truthful feedback for teachers while paving the way for future research to address these obstacles and improve the capacity of generative AI to coach teachers.

We present the B-fields mapped in IRDC G34.43+0.24 using 850\,μm polarized dust emission observed with the POL-2 instrument at JCMT. We examine the magnetic field geometries and strengths in the northern, central, and southern regions of the filament. The overall field geometry is ordered and aligned closely perpendicular to the filament's main axis, particularly in regions containing the central clumps MM1 and MM2, whereas MM3 in the north has field orientations aligned with its major axis. The overall field orientations are uniform at large (POL-2 at 14arcsec and SHARP at 10arcsec) to small scales (TADPOL at 2.5arcsec and SMA at 1.5arcsec) in the MM1 and MM2 regions. SHARP/CSO observations in MM3 at 350\,μm from Tang et al. show a similar trend as seen in our POL-2 observations. TADPOL observations demonstrate a well-defined field geometry in MM1/MM2 consistent with MHD simulations of accreting filaments. We obtained a plane-of-sky magnetic field strength of 470pm190\,μG, 100pm40\,μG, and 60pm34\,μG in the central, northern and southern regions of G34, respectively, using the updated Davis-Chandrasekhar-Fermi relation. The estimated value of field strength, combined with column density and velocity dispersion values available in the literature, suggests G34 to be marginally critical with criticality parameter rm λ values 0.8pm0.4, 1.1pm0.8, and 0.9pm0.5 in the central, northern, and southern regions, respectively. The turbulent motions in G34 are sub-Alfvénic with Alfvénic Mach numbers of 0.34pm0.13, 0.53pm0.30, and 0.49pm0.26 in the three regions. The observed aligned B-fields in G34.43+0.24 are consistent with theoretical models suggesting that B-fields play an important role in guiding the contraction of the cloud driven by gravity.

As opposed to evaluating computation and logic-based reasoning, current benchmarks for evaluating large language models (LLMs) in medicine are primarily focused on question-answering involving domain knowledge and descriptive reasoning. While such qualitative capabilities are vital to medical diagnosis, in real-world scenarios, doctors frequently use clinical calculators that follow quantitative equations and rule-based reasoning paradigms for evidence-based decision support. To this end, we propose MedCalc-Bench, a first-of-its-kind dataset focused on evaluating the medical calculation capability of LLMs. MedCalc-Bench contains an evaluation set of over 1000 manually reviewed instances from 55 different medical calculation tasks. Each instance in MedCalc-Bench consists of a patient note, a question requesting to compute a specific medical value, a ground truth answer, and a step-by-step explanation showing how the answer is obtained. While our evaluation results show the potential of LLMs in this area, none of them are effective enough for clinical settings. Common issues include extracting the incorrect entities, not using the correct equation or rules for a calculation task, or incorrectly performing the arithmetic for the computation. We hope our study highlights the quantitative knowledge and reasoning gaps in LLMs within medical settings, encouraging future improvements of LLMs for various clinical calculation tasks.

Modern astronomical surveys, such as the Euclid mission, produce high-dimensional, multi-modal data sets that include imaging and spectroscopic information for millions of galaxies. These data serve as an ideal benchmark for large, pre-trained multi-modal models, which can leverage vast amounts of unlabelled data. In this work, we present the first exploration of Euclid data with AstroPT, an autoregressive multi-modal foundation model trained on approximately 300 000 optical and infrared Euclid images and spectral energy distributions (SEDs) from the first Euclid Quick Data Release. We compare self-supervised pre-training with baseline fully supervised training across several tasks: galaxy morphology classification; redshift estimation; similarity searches; and outlier detection. Our results show that: (a) AstroPT embeddings are highly informative, correlating with morphology and effectively isolating outliers; (b) including infrared data helps to isolate stars, but degrades the identification of edge-on galaxies, which are better captured by optical images; (c) simple fine-tuning of these embeddings for photometric redshift and stellar mass estimation outperforms a fully supervised approach, even when using only 1% of the training labels; and (d) incorporating SED data into AstroPT via a straightforward multi-modal token-chaining method improves photo-z predictions, and allow us to identify potentially more interesting anomalies (such as ringed or interacting galaxies) compared to a model pre-trained solely on imaging data.

This document describes the technical layout and the expected performance of the Straw Tube Tracker (STT), the main tracking detector of the PANDA target spectrometer. The STT encloses a Micro-Vertex-Detector (MVD) for the inner tracking and is followed in beam direction by a set of GEM-stations. The tasks of the STT are the measurement of the particle momentum from the reconstructed trajectory and the measurement of the specific energy-loss for a particle identification. Dedicated simulations with full analysis studies of certain proton-antiproton reactions, identified as being benchmark tests for the whole PANDA scientific program, have been performed to test the STT layout and performance. The results are presented, and the time lines to construct the STT are described.

Large language models (LLMs) are rapidly changing how researchers in materials science and chemistry discover, organize, and act on scientific knowledge. This paper analyzes a broad set of community-developed LLM applications in an effort to identify emerging patterns in how these systems can be used across the scientific research lifecycle. We organize the projects into two complementary categories: Knowledge Infrastructure, systems that structure, retrieve, synthesize, and validate scientific information; and Action Systems, systems that execute, coordinate, or automate scientific work across computational and experimental environments. The submissions reveal a shift from single-purpose LLM tools toward integrated, multi-agent workflows that combine retrieval, reasoning, tool use, and domain-specific validation. Prominent themes include retrieval-augmented generation as grounding infrastructure, persistent structured knowledge representations, multimodal and multilingual scientific inputs, and early progress toward laboratory-integrated closed-loop systems. Together, these results suggest that LLMs are evolving from general-purpose assistants into composable infrastructure for scientific reasoning and action. This work provides a community snapshot of that transition and a practical taxonomy for understanding emerging LLM-enabled workflows in materials science and chemistry.

We show, for the first time, radio measurements of the depth of shower maximum (X_max) of air showers induced by cosmic rays that are compared to measurements of the established fluorescence method at the same location. Using measurements at the Pierre Auger Observatory we show full compatibility between our radio and the previously published fluorescence data set, and between a subset of air showers observed simultaneously with both radio and fluorescence techniques, a measurement setup unique to the Pierre Auger Observatory. Furthermore, we show radio X_max resolution as a function of energy and demonstrate the ability to make competitive high-resolution X_max measurements with even a sparse radio array. With this, we show that the radio technique is capable of cosmic-ray mass composition studies, both at Auger and at other experiments.

Data augmentation is an important component in the robustness evaluation of models in natural language processing (NLP) and in enhancing the diversity of the data they are trained on. In this paper, we present NL-Augmenter, a new participatory Python-based natural language augmentation framework which supports the creation of both transformations (modifications to the data) and filters (data splits according to specific features). We describe the framework and an initial set of 117 transformations and 23 filters for a variety of natural language tasks. We demonstrate the efficacy of NL-Augmenter by using several of its transformations to analyze the robustness of popular natural language models. The infrastructure, datacards and robustness analysis results are available publicly on the NL-Augmenter repository (https://github.com/GEM-benchmark/NL-Augmenter).

Euclid is an ESA mission designed to constrain the properties of dark energy and gravity via weak gravitational lensing and galaxy clustering. It will carry out a wide area imaging and spectroscopy survey (EWS) in visible and near-infrared, covering roughly 15,000 square degrees of extragalactic sky on six years. The wide-field telescope and instruments are optimized for pristine PSF and reduced straylight, producing very crisp images. This paper presents the building of the Euclid reference survey: the sequence of pointings of EWS, Deep fields, Auxiliary fields for calibrations, and spacecraft movements followed by Euclid as it operates in a step-and-stare mode from its orbit around the Lagrange point L2. Each EWS pointing has four dithered frames; we simulate the dither pattern at pixel level to analyse the effective coverage. We use up-to-date models for the sky background to define the Euclid region-of-interest (RoI). The building of the reference survey is highly constrained from calibration cadences, spacecraft constraints and background levels; synergies with ground-based coverage are also considered. Via purposely-built software optimized to prioritize best sky areas, produce a compact coverage, and ensure thermal stability, we generate a schedule for the Auxiliary and Deep fields observations and schedule the RoI with EWS transit observations. The resulting reference survey RSD_2021A fulfills all constraints and is a good proxy for the final solution. Its wide survey covers 14,500 square degrees. The limiting AB magnitudes (5σ point-like source) achieved in its footprint are estimated to be 26.2 (visible) and 24.5 (near-infrared); for spectroscopy, the H_α line flux limit is 2times 10^{-16} erg cm^{-2} s^{-1} at 1600 nm; and for diffuse emission the surface brightness limits are 29.8 (visible) and 28.4 (near-infrared) mag arcsec^{-2}.

We report on Quantinuum Helios, a 98-qubit trapped-ion quantum processor based on the quantum charge-coupled device (QCCD) architecture. Helios features ^{137}Ba^{+} hyperfine qubits, all-to-all connectivity enabled by a rotatable ion storage ring connecting two quantum operation regions by a junction, speed improvements from parallelized operations, and a new software stack with real-time compilation of dynamic programs. Averaged over all operational zones in the system, we achieve average infidelities of 2.5(1)times10^{-5} for single-qubit gates, 7.9(2)times10^{-4} for two-qubit gates, and 4.8(6)times10^{-4} for state preparation and measurement, none of which are fundamentally limited and likely able to be improved. These component infidelities are predictive of system-level performance in both random Clifford circuits and random circuit sampling, the latter demonstrating that Helios operates well beyond the reach of classical simulation and establishes a new frontier of fidelity and complexity for quantum computers.

We present a detailed visual morphology catalogue for Euclid's Quick Release 1 (Q1). Our catalogue includes galaxy features such as bars, spiral arms, and ongoing mergers, for the 378000 bright (IE < 20.5) or extended (area geq 700,pixels) galaxies in Q1. The catalogue was created by finetuning the Zoobot galaxy foundation models on annotations from an intensive one month campaign by Galaxy Zoo volunteers. Our measurements are fully automated and hence fully scaleable. This catalogue is the first 0.4% of the approximately 100 million galaxies where Euclid will ultimately resolve detailed morphology.

The Euclid satellite is an ESA mission that was launched in July 2023. \Euclid is working in its regular observing mode with the target of observing an area of 14,000~deg^2 with two instruments, the Visible Camera (VIS) and the Near IR Spectrometer and Photometer (NISP) down to I_{rm E} = 24.5~mag (10, sigma) in the Euclid Wide Survey. Ground-based imaging data in the ugriz bands complement the \Euclid data to enable photo-z determination and VIS PSF modeling for week lensing analysis. Euclid investigates the distance-redshift relation and the evolution of cosmic structures by measuring shapes and redshifts of galaxies and clusters of galaxies out to zsim 2. Generating the multi-wavelength catalogues from \Euclid and ground-based data is an essential part of the \Euclid data processing system. In the framework of the \Euclid Science Ground Segment (SGS), the aim of the MER Processing Function (PF) pipeline is to detect objects in the \Euclid imaging data, measure their properties, and MERge them into a single multi-wavelength catalogue. The MER PF pipeline performs source detection on both visible (VIS) and near-infrared (NIR) images and offers four different photometric measurements: Kron total flux, aperture photometry on PSF-matched images, template fitting photometry, and S\'ersic fitting photometry. Furthermore, the MER PF pipeline measures a set of ancillary quantities, spanning from morphology to quality flags, to better characterise all detected sources. In this paper, we show how the MER PF pipeline is designed, detailing its main steps, and we show that the pipeline products meet the tight requirements that Euclid aims to achieve on photometric accuracy. We also present the other measurements (e.g. morphology) that are included in the OU-MER output catalogues and we list all output products coming out of the MER PF pipeline.

The current standard model of cosmology successfully describes a variety of measurements, but the nature of its main ingredients, dark matter and dark energy, remains unknown. Euclid is a medium-class mission in the Cosmic Vision 2015-2025 programme of the European Space Agency (ESA) that will provide high-resolution optical imaging, as well as near-infrared imaging and spectroscopy, over about 14,000 deg^2 of extragalactic sky. In addition to accurate weak lensing and clustering measurements that probe structure formation over half of the age of the Universe, its primary probes for cosmology, these exquisite data will enable a wide range of science. This paper provides a high-level overview of the mission, summarising the survey characteristics, the various data-processing steps, and data products. We also highlight the main science objectives and expected performance.

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai.

Memory is critical for AI agents, yet the widely-adopted static memory, aiming to create readily available memory in advance, is inevitably subject to severe information loss. To address this limitation, we propose a novel framework called general agentic memory (GAM). GAM follows the principle of "just-in time (JIT) compilation" where it focuses on creating optimized contexts for its client at runtime while keeping only simple but useful memory during the offline stage. To this end, GAM employs a duo-design with the following components. 1) Memorizer, which highlights key historical information using a lightweight memory, while maintaining complete historical information within a universal page-store. 2) Researcher, which retrieves and integrates useful information from the page-store for its online request guided by the pre-constructed memory. This design allows GAM to effectively leverage the agentic capabilities and test-time scalability of frontier large language models (LLMs), while also facilitating end-to-end performance optimization through reinforcement learning. In our experimental study, we demonstrate that GAM achieves substantial improvement on various memory-grounded task completion scenarios against existing memory systems.

Reasoning benchmarks have played a crucial role in the progress of language models. Yet rigorous evaluation remains a significant challenge as static question-answer pairs provide only a snapshot of performance, compressing complex behavior into a single accuracy metric. This limitation is especially true in complex, rule-bound domains such as law, where existing benchmarks are costly to build and ill suited for isolating specific failure modes. To address this, we introduce OpenExempt, a framework and benchmark for diagnostic evaluation of legal reasoning. The OpenExempt Framework uses expert-crafted symbolic representations of U.S. Bankruptcy Code statutes to dynamically generate a large space of natural language reasoning tasks and their machine-computable solutions on demand. This gives users fine-grained control over task complexity and scope, allowing individual reasoning skills to be probed in isolation. Using this system, we construct the OpenExempt Benchmark, a diagnostic benchmark for legal reasoning with 9,765 samples across nine evaluation suites designed to carefully probe model capabilities. Experiments on 13 diverse language models reveal sharp performance cliffs that emerge only under longer reasoning paths and in the presence of obfuscating statements. We release the framework and benchmark publicly to support research aimed at understanding and improving the next generation of reasoning systems.

Statistical language modeling and translation with transformers have found many successful applications in program understanding and generation tasks, setting high benchmarks for tools in modern software development environments. The finite context window of these neural models means, however, that they will be unable to leverage the entire relevant context of large files and packages for any given task. While there are many efforts to extend the context window, we introduce an architecture-independent approach for leveraging the syntactic hierarchies of source code for incorporating entire file-level context into a fixed-length window. Using concrete syntax trees of each source file we extract syntactic hierarchies and integrate them into context window by selectively removing from view more specific, less relevant scopes for a given task. We evaluate this approach on code generation tasks and joint translation of natural language and source code in Python programming language, achieving a new state-of-the-art in code completion and summarization for Python in the CodeXGLUE benchmark. We also introduce new CodeXGLUE benchmarks for user-experience-motivated tasks: code completion with normalized literals, method body completion/code summarization conditioned on file-level context.

This paper is an initial endeavor to bridge the gap between powerful Deep Reinforcement Learning methodologies and the problem of exploration/coverage of unknown terrains. Within this scope, MarsExplorer, an openai-gym compatible environment tailored to exploration/coverage of unknown areas, is presented. MarsExplorer translates the original robotics problem into a Reinforcement Learning setup that various off-the-shelf algorithms can tackle. Any learned policy can be straightforwardly applied to a robotic platform without an elaborate simulation model of the robot's dynamics to apply a different learning/adaptation phase. One of its core features is the controllable multi-dimensional procedural generation of terrains, which is the key for producing policies with strong generalization capabilities. Four different state-of-the-art RL algorithms (A3C, PPO, Rainbow, and SAC) are trained on the MarsExplorer environment, and a proper evaluation of their results compared to the average human-level performance is reported. In the follow-up experimental analysis, the effect of the multi-dimensional difficulty setting on the learning capabilities of the best-performing algorithm (PPO) is analyzed. A milestone result is the generation of an exploration policy that follows the Hilbert curve without providing this information to the environment or rewarding directly or indirectly Hilbert-curve-like trajectories. The experimental analysis is concluded by evaluating PPO learned policy algorithm side-by-side with frontier-based exploration strategies. A study on the performance curves revealed that PPO-based policy was capable of performing adaptive-to-the-unknown-terrain sweeping without leaving expensive-to-revisit areas uncovered, underlying the capability of RL-based methodologies to tackle exploration tasks efficiently. The source code can be found at: https://github.com/dimikout3/MarsExplorer.

PennyLane is a Python 3 software framework for differentiable programming of quantum computers. The library provides a unified architecture for near-term quantum computing devices, supporting both qubit and continuous-variable paradigms. PennyLane's core feature is the ability to compute gradients of variational quantum circuits in a way that is compatible with classical techniques such as backpropagation. PennyLane thus extends the automatic differentiation algorithms common in optimization and machine learning to include quantum and hybrid computations. A plugin system makes the framework compatible with any gate-based quantum simulator or hardware. We provide plugins for hardware providers including the Xanadu Cloud, Amazon Braket, and IBM Quantum, allowing PennyLane optimizations to be run on publicly accessible quantum devices. On the classical front, PennyLane interfaces with accelerated machine learning libraries such as TensorFlow, PyTorch, JAX, and Autograd. PennyLane can be used for the optimization of variational quantum eigensolvers, quantum approximate optimization, quantum machine learning models, and many other applications.

The lack of large and diverse training data on Computer-Aided Diagnosis (CAD) in breast cancer detection has been one of the concerns that impedes the adoption of the system. Recently, pre-training with large-scale image text datasets via Vision-Language models (VLM) (\eg CLIP) partially addresses the issue of robustness and data efficiency in computer vision (CV). This paper proposes Mammo-CLIP, the first VLM pre-trained on a substantial amount of screening mammogram-report pairs, addressing the challenges of dataset diversity and size. Our experiments on two public datasets demonstrate strong performance in classifying and localizing various mammographic attributes crucial for breast cancer detection, showcasing data efficiency and robustness similar to CLIP in CV. We also propose Mammo-FActOR, a novel feature attribution method, to provide spatial interpretation of representation with sentence-level granularity within mammography reports. Code is available publicly: https://github.com/batmanlab/Mammo-CLIP.

The study of twisted two-dimensional (2D) materials, where twisting layers create moiré superlattices, has opened new opportunities for investigating topological phases and strongly correlated physics. While systems such as twisted bilayer graphene (TBG) and twisted transition metal dichalcogenides (TMDs) have been extensively studied, the broader potential of a seemingly infinite set of other twistable 2D materials remains largely unexplored. In this paper, we define "theoretically twistable materials" as single- or multi-layer structures that allow for the construction of simple continuum models of their moiré structures. This excludes, for example, materials with a "spaghetti" of bands or those with numerous crossing points at the Fermi level, for which theoretical moiré modeling is unfeasible. We present a high-throughput algorithm that systematically searches for theoretically twistable semimetals and insulators based on the Topological 2D Materials Database. By analyzing key electronic properties, we identify thousands of new candidate materials that could host rich topological and strongly correlated phenomena when twisted. We propose representative twistable materials for realizing different types of moiré systems, including materials with different Bravais lattices, valleys, and strength of spin-orbital coupling. We provide examples of crystal growth for several of these materials and showcase twisted bilayer band structures along with simplified twisted continuum models. Our results significantly broaden the scope of moiré heterostructures and provide a valuable resource for future experimental and theoretical studies on novel moiré systems.