Serena Yeung-Levy

Laura Bravo-Sánchez, Matthieu Armando, Romain Brégier et al.

Recovering 3D human pose and shape from a single image remains a cornerstone of human-centric vision, yet most methods assume adult subjects and optimize each person independently. These assumptions fail in real-world, all-age scenes, where body proportions and depth must be resolved jointly. We introduce Anny-Fit, a multi-person, camera-space optimization framework for all-age 3D human mesh recovery (HMR). Unlike existing per-person fitting methods, Anny-Fit jointly optimizes all individuals directly in the camera coordinate system, enforcing global spatial consistency. At the core of our approach is the use of multiple forms of expert knowledge -- including metric depth maps, instance segmentation, 2D keypoints, and, VLM-derived semantic attributes such as age and gender -- each obtained from dedicated off-the-shelf networks. These complementary signals jointly guide the optimization, constraining the depth-scale ambiguity characteristic of all-age scenes. Across diverse datasets, Anny-Fit consistently improves 2D reprojection accuracy (+13 to 16), relative depth ordering (+6 to 7), 3D estimation error (-9 to -29) and shape estimation (+25 to +82), producing more coherent scenes. Finally, we show that VLM-based semantic knowledge can be distilled into an HMR model via the pseudo-ground-truth annotations produced by Anny-Fit on training data, enabling it to learn semantically meaningful shape parameters while improving HMR performance. Our approach bridges adult-only and all-age modeling by enabling zero-shot adaptation of adult-trained HMR pipelines to the full age spectrum without retraining. Code is publicly available at https://github.com/naver/anny-fit.

Charlotte Bunne, Yusuf Roohani, Yanay Rosen et al.

The cell is arguably the most fundamental unit of life and is central to understanding biology. Accurate modeling of cells is important for this understanding as well as for determining the root causes of disease. Recent advances in artificial intelligence (AI), combined with the ability to generate large-scale experimental data, present novel opportunities to model cells. Here we propose a vision of leveraging advances in AI to construct virtual cells, high-fidelity simulations of cells and cellular systems under different conditions that are directly learned from biological data across measurements and scales. We discuss desired capabilities of such AI Virtual Cells, including generating universal representations of biological entities across scales, and facilitating interpretable in silico experiments to predict and understand their behavior using virtual instruments. We further address the challenges, opportunities and requirements to realize this vision including data needs, evaluation strategies, and community standards and engagement to ensure biological accuracy and broad utility. We envision a future where AI Virtual Cells help identify new drug targets, predict cellular responses to perturbations, as well as scale hypothesis exploration. With open science collaborations across the biomedical ecosystem that includes academia, philanthropy, and the biopharma and AI industries, a comprehensive predictive understanding of cell mechanisms and interactions has come into reach.

Alejandro Lozano, Keiko Ihara, Ping-Hao Yang et al.

Summarizing the latest medical literature to guide clinical decision-making is essential for evidence-based medicine and high-quality patient care. Yet clinicians face increasing challenges due to limited time with patients and a rapidly growing volume of published articles. Although retrieval-augmented large language models (LLMs) have shown promise in clinical summarization, human evaluations of their effectiveness in synthesizing broader scientific literature and direct comparisons to expert-written syntheses remain scarce. We constructed a RAG-based agentic AI framework using three state-of-the-art LLMs: Sonnet, GPT-4o, and Llama 3.1. A headache specialist created 13 questions, three for prompt optimization and ten for evaluation. Ten headache specialists across the United States and Canada each wrote a summary for one question, yielding four summaries per question (expert, Sonnet, GPT-4o, and Llama). The experts, blinded to authorship, critically evaluated the summaries, excluding the topic for which they wrote a summary, based on correctness, completeness, conciseness, and clinical utility, scoring each from 1 to 10 using standardized rubrics. They also ranked the summaries by preference and indicated whether they believed each summary was written by an expert or an LLM. Our study, comparing LLM- and expert-written literature summaries evaluated by headache specialists, showed that expert-written summaries were preferred, although experts sometimes found it challenging to distinguish between human- and AI-generated summaries. We also identified key expert-valued features beyond standard evaluation metrics that can guide future refinement of both human and AI literature summarization pipelines.

Junzhe Huang, Xiaoxiao Sun, Yan Yang et al.

Using multimodal foundation models to analyze table images is a high-value yet challenging application in consumer and enterprise scenarios. Despite its importance, current evaluations rely largely on structured-text tables or clean rendered images, leaving the visual complexity of in-the-wild table images underexplored. Such images feature varied layouts and diverse domains that demand sophisticated structural perception and numerical reasoning. To bridge this gap, we introduce WildTableBench, the first question-answering benchmark for naturally occurring table images from real-world settings. WildTableBench comprises 402 high-information-density table images collected from online forums and websites across diverse domains, together with 928 manually annotated and verified questions spanning 17 subtypes across five categories. We evaluate 21 frontier proprietary and open-source multimodal foundation models on this benchmark. Only one model exceeds 50% accuracy, while all remaining models range from 4.1% to 49.9%. We further conduct diagnostic analyses to characterize model failures and reveal persistent weaknesses in structural perception and reasoning. These results and analyses provide useful insights into current model capabilities and establish WildTableBench as a valuable diagnostic benchmark for table image understanding.

Alejandro Lozano, Jeffrey Nirschl, James Burgess et al. · stanford

Recent advances in microscopy have enabled the rapid generation of terabytes of image data in cell biology and biomedical research. Vision-language models (VLMs) offer a promising solution for large-scale biological image analysis, enhancing researchers' efficiency, identifying new image biomarkers, and accelerating hypothesis generation and scientific discovery. However, there is a lack of standardized, diverse, and large-scale vision-language benchmarks to evaluate VLMs' perception and cognition capabilities in biological image understanding. To address this gap, we introduce μ-Bench, an expert-curated benchmark encompassing 22 biomedical tasks across various scientific disciplines (biology, pathology), microscopy modalities (electron, fluorescence, light), scales (subcellular, cellular, tissue), and organisms in both normal and abnormal states. We evaluate state-of-the-art biomedical, pathology, and general VLMs on μ-Bench and find that: i) current models struggle on all categories, even for basic tasks such as distinguishing microscopy modalities; ii) current specialist models fine-tuned on biomedical data often perform worse than generalist models; iii) fine-tuning in specific microscopy domains can cause catastrophic forgetting, eroding prior biomedical knowledge encoded in their base model. iv) weight interpolation between fine-tuned and pre-trained models offers one solution to forgetting and improves general performance across biomedical tasks. We release μ-Bench under a permissive license to accelerate the research and development of microscopy foundation models.

James Burgess, Kuan-Chieh Wang, Serena Yeung-Levy

Text-to-image diffusion models generate impressive and realistic images, but do they learn to represent the 3D world from only 2D supervision? We demonstrate that yes, certain 3D scene representations are encoded in the text embedding space of models like Stable Diffusion. Our approach, Viewpoint Neural Textual Inversion (ViewNeTI), is to discover 3D view tokens; these tokens control the 3D viewpoint - the rendering pose in a scene - of generated images. Specifically, we train a small neural mapper to take continuous camera viewpoint parameters and predict a view token (a word embedding). This token conditions diffusion generation via cross-attention to produce images with the desired camera viewpoint. Using ViewNeTI as an evaluation tool, we report two findings: first, the text latent space has a continuous view-control manifold for particular 3D scenes; second, we find evidence for a generalized view-control manifold for all scenes. We conclude that since the view token controls the 3D `rendering' viewpoint, there is likely a scene representation embedded in frozen 2D diffusion models. Finally, we exploit the 3D scene representations for 3D vision tasks, namely, view-controlled text-to-image generation, and novel view synthesis from a single image, where our approach sets state-of-the-art for LPIPS. Code available at https://github.com/jmhb0/view_neti

Zhongying Deng, Cheng Tang, Ziyan Huang et al. · pku

Foundation models have demonstrated remarkable success across diverse domains and tasks, primarily due to the thrive of large-scale, diverse, and high-quality datasets. However, in the field of medical imaging, the curation and assembling of such medical datasets are highly challenging due to the reliance on clinical expertise and strict ethical and privacy constraints, resulting in a scarcity of large-scale unified medical datasets and hindering the development of powerful medical foundation models. In this work, we present the largest survey to date of medical image datasets, covering over 1,000 open-access datasets with a systematic catalog of their modalities, tasks, anatomies, annotations, limitations, and potential for integration. Our analysis exposes a landscape that is modest in scale, fragmented across narrowly scoped tasks, and unevenly distributed across organs and modalities, which in turn limits the utility of existing medical image datasets for developing versatile and robust medical foundation models. To turn fragmentation into scale, we propose a metadata-driven fusion paradigm (MDFP) that integrates public datasets with shared modalities or tasks, thereby transforming multiple small data silos into larger, more coherent resources. Building on MDFP, we release an interactive discovery portal that enables end-to-end, automated medical image dataset integration, and compile all surveyed datasets into a unified, structured table that clearly summarizes their key characteristics and provides reference links, offering the community an accessible and comprehensive repository. By charting the current terrain and offering a principled path to dataset consolidation, our survey provides a practical roadmap for scaling medical imaging corpora, supporting faster data discovery, more principled dataset creation, and more capable medical foundation models.

Wenjin Hou, Wei Liu, Han Hu et al.

Multimodal Large Language Models (MLLMs) have shown remarkable proficiency on general-purpose vision-language benchmarks, reaching or even exceeding human-level performance. However, these evaluations typically rely on standard in-distribution data, leaving the robustness of MLLMs largely unexamined when faced with scenarios that defy common-sense priors. To address this gap, we introduce VIA-Bench, a challenging benchmark designed to probe model performance on visual illusions and anomalies. It includes six core categories: color illusions, motion illusions, gestalt illusions, geometric and spatial illusions, general visual illusions, and visual anomalies. Through careful human-in-the-loop review, we construct over 1K high-quality question-answer pairs that require nuanced visual reasoning. Extensive evaluation of over 20 state-of-the-art MLLMs, including proprietary, open-source, and reasoning-enhanced models, uncovers significant vulnerabilities. Notably, we find that Chain-of-Thought (CoT) reasoning offers negligible robustness, often yielding ``brittle mirages'' where the model's logic collapses under illusory stimuli. Our findings reveal a fundamental divergence between machine and human perception, suggesting that resolving such perceptual bottlenecks is critical for the advancement of artificial general intelligence. The benchmark data and code will be released.

James Burgess, Jan N. Hansen, Duo Peng et al.

Search agents are language models (LMs) that reason and search knowledge bases (or the web) to answer questions; recent methods supervise only the final answer accuracy using reinforcement learning with verifiable rewards (RLVR). Most RLVR search agents tackle general-domain QA, which limits their relevance to technical AI systems in science, engineering, and medicine. In this work we propose training agents to search and reason over scientific papers -- this tests technical question-answering, it is directly relevant to real scientists, and the capabilities will be crucial to future AI Scientist systems. Concretely, we release a search corpus of 16 million biomedical paper abstracts and construct a challenging factoid QA dataset called PaperSearchQA with 60k samples answerable from the corpus, along with benchmarks. We train search agents in this environment to outperform non-RL retrieval baselines; we also perform further quantitative analysis and observe interesting agent behaviors like planning, reasoning, and self-verification. Our corpus, datasets, and benchmarks are usable with the popular Search-R1 codebase for RLVR training and released on https://huggingface.co/collections/jmhb/papersearchqa. Finally, our data creation methods are scalable and easily extendable to other scientific domains.

Orr Zohar, Xiaohan Wang, Yonatan Bitton et al.

The performance of Large Vision Language Models (LVLMs) is dependent on the size and quality of their training datasets. Existing video instruction tuning datasets lack diversity as they are derived by prompting large language models with video captions to generate question-answer pairs, and are therefore mostly descriptive. Meanwhile, many labeled video datasets with diverse labels and supervision exist - however, we find that their integration into LVLMs is non-trivial. Herein, we present Video Self-Training with augmented Reasoning (Video-STaR), the first video self-training approach. Video-STaR allows the utilization of any labeled video dataset for video instruction tuning. In Video-STaR, an LVLM cycles between instruction generation and finetuning, which we show (I) improves general video understanding and (II) adapts LVLMs to novel downstream tasks with existing supervision. During generation, an LVLM is prompted to propose an answer. The answers are then filtered only to those that contain the original video labels, and the LVLM is then re-trained on the generated dataset. By only training on generated answers that contain the correct video labels, Video-STaR utilizes these existing video labels as weak supervision for video instruction tuning. Our results demonstrate that Video-STaR-enhanced LVLMs exhibit improved performance in (I) general video QA, where TempCompass performance improved by 10%, and (II) on downstream tasks, where Video-STaR improved Kinetics700-QA accuracy by 20% and action quality assessment on FineDiving by 15%.

Zeyu Wang, Zhenzhen Weng, Serena Yeung-Levy

Continuous perception, the ability to integrate visual observations over time in a continuous stream fashion, is essential for robust real-world understanding, yet remains largely untested in current multimodal models. We introduce CP-Bench, a minimal and fully controlled benchmark designed to isolate this capability using an extremely simple task: counting identical cubes in a synthetic scene while the camera moves and only reveals subsets of objects at any moment. Despite the simplicity of the setting, we find that state-of-the-art open-source and commercial models, including Qwen-3-VL, InternVL3, GPT-5, and Gemini-3-Pro, fail dramatically. A static-camera control variant confirms that the failure arises not from object recognition but from an inability to accumulate evidence across time. Further experiments show that neither higher sampling FPS, perception- or spatial-enhanced models, nor finetuning with additional videos leads to meaningful cross-temporal generalization. Our results reveal a fundamental limitation in modern multimodal architectures and training paradigms. CP-Bench provides a simple yet powerful diagnostic tool and establishes a clean testbed for developing models capable of genuine time-consistent visual reasoning.

Zhenzhen Weng, Laura Bravo-Sánchez, Serena Yeung-Levy

Recent text-to-image generative models have exhibited remarkable abilities in generating high-fidelity and photo-realistic images. However, despite the visually impressive results, these models often struggle to preserve plausible human structure in the generations. Due to this reason, while generative models have shown promising results in aiding downstream image recognition tasks by generating large volumes of synthetic data, they are not suitable for improving downstream human pose perception and understanding. In this work, we propose a Diffusion model with Human Pose Correction (Diffusion-HPC), a text-conditioned method that generates photo-realistic images with plausible posed humans by injecting prior knowledge about human body structure. Our generated images are accompanied by 3D meshes that serve as ground truths for improving Human Mesh Recovery tasks, where a shortage of 3D training data has long been an issue. Furthermore, we show that Diffusion-HPC effectively improves the realism of human generations under varying conditioning strategies.

Dongxia Wu, Yuhui Zhang, Serena Yeung-Levy et al.

Distribution-to-distribution generative models support scientific imaging tasks ranging from modeling cellular perturbation responses to translating medical images across conditions. Trustworthy generation requires both reliability (generalization across labs, devices, and experimental conditions) and accountability (detecting out-of-distribution cases where predictions may be unreliable). Uncertainty quantification (UQ) based approaches serve as promising candidates for these tasks, yet UQ for distribution-to-distribution generative models remains underexplored. We present a unified UQ framework, Bayesian Stochastic Flow Matching (BSFM), that disentangles aleatoric and epistemic uncertainty. The Stochastic Flow Matching (SFM) component augments deterministic flows with a diffusion term to improve model generalization to unseen scenarios. For UQ, we develop a scalable Bayesian approach -- MCD-Antithetic -- that combines Monte Carlo Dropout with sample-efficient antithetic sampling to produce effective anomaly scores for out-of-distribution detection. Experiments on cellular imaging (BBBC021, JUMP) and brain fMRI (Theory of Mind) across diverse scenarios show that SFM improves reliability while MCD-Antithetic enhances accountability.

Xiaoxiao Sun, Mingyang Li, Kun yuan et al.

Large Vision-Language Models (VLMs) often answer classic visual illusions "correctly" on original images, yet persist with the same responses when illusion factors are inverted, even though the visual change is obvious to humans. This raises a fundamental question: do VLMs perceive visual changes or merely recall memorized patterns? While several studies have noted this phenomenon, the underlying causes remain unclear. To move from observations to systematic understanding, this paper introduces VI-Probe, a controllable visual-illusion framework with graded perturbations and matched visual controls (without illusion inducer) that disentangles visually grounded perception from language-driven recall. Unlike prior work that focuses on averaged accuracy, we measure stability and sensitivity using Polarity-Flip Consistency, Template Fixation Index, and an illusion multiplier normalized against matched controls. Experiments across different families reveal that response persistence arises from heterogeneous causes rather than a single mechanism. For instance, GPT-5 exhibits memory override, Claude-Opus-4.1 shows perception-memory competition, while Qwen variants suggest visual-processing limits. Our findings challenge single-cause views and motivate probing-based evaluation that measures both knowledge and sensitivity to controlled visual change. Data and code are available at https://sites.google.com/view/vi-probe/.

James Burgess, Rameen Abdal, Dan Stoddart et al.

Modern image generators produce strikingly realistic images, where only artifacts like distorted hands or warped objects reveal their synthetic origin. Detecting these artifacts is essential: without detection, we cannot benchmark generators or train reward models to improve them. Current detectors fine-tune VLMs on tens of thousands of labeled images, but this is expensive to repeat whenever generators evolve or new artifact types emerge. We show that pretrained VLMs already encode the knowledge needed to detect artifacts - with the right scaffolding, this capability can be unlocked using only a few hundred labeled examples per artifact category. Our system, ArtifactLens, achieves state-of-the-art on five human artifact benchmarks (the first evaluation across multiple datasets) while requiring orders of magnitude less labeled data. The scaffolding consists of a multi-component architecture with in-context learning and text instruction optimization, with novel improvements to each. Our methods generalize to other artifact types - object morphology, animal anatomy, and entity interactions - and to the distinct task of AIGC detection.

Ruotong Liao, Nikolai Röhrich, Xiaohan Wang et al.

Test-time reinforcement learning (TTRL) has emerged as a promising paradigm for self-evolving large reasoning models (LRMs), enabling online adaptation on unlabeled test inputs via self-induced rewards through majority voting. However, a spurious yet high-frequency unverified consensus can become a biased and reinforced reward signal, leading to incorrect mode collapse. We address this failure mode with T^3RL (Tool-Verification for Test-Time Reinforcement Learning), which introduces test-time tool verification into reward estimation. Concretely, a verifier uses an external tool as evidence (e.g., from code execution) to upweight verified rollouts in a verification-aware voting, producing more reliable pseudo-labels for training. Across various math difficulties (MATH-500, AMC, and AIME 2024) and diverse backbone types, T^3RL significantly improves over TTRL, with larger gains on harder problems. More broadly, T^3RL can be viewed as verified online data synthesis, highlighting test-time tool verification as a key mechanism for stabilizing self-evolution.

Shengguang Wu, Xiaohan Wang, Yuhui Zhang et al.

Spatial reasoning in 3D scenes requires precise geometric calculations that challenge vision-language models. Visual programming addresses this by decomposing problems into steps calling specialized tools, yet existing methods rely on either fixed toolsets or speculative tool induction before solving problems, resulting in suboptimal programs and poor utilization of induced tools. We present Transductive Visual Programming (TVP), a novel framework that builds new tools from its own experience rather than speculation. TVP first solves problems using basic tools while accumulating experiential solutions into an Example Library, then abstracts recurring patterns from these programs into reusable higher-level tools for an evolving Tool Library. This allows TVP to tackle new problems with increasingly powerful tools learned from experience. On Omni3D-Bench, TVP achieves state-of-the-art performance, outperforming GPT-4o by 22% and the previous best visual programming system by 11%. Our transductively learned tools are used 5x more frequently as core program dependency than inductively created ones, demonstrating more effective tool discovery and reuse. The evolved tools also show strong generalization to unseen spatial tasks, achieving superior performance on benchmarks from SpatialScore-Hard collection without any testset-specific modification. Our work establishes experience-driven transductive tool creation as a powerful paradigm for building self-evolving visual programming agents that effectively tackle challenging spatial reasoning tasks. We release our code at https://transductive-visualprogram.github.io/.

Dongxia Wu, Shiye Su, Yuhui Zhang et al.

Building virtual cells with generative models to simulate cellular behavior in silico is emerging as a promising paradigm for accelerating drug discovery. However, prior image-based generative approaches can produce implausible cell images that violate basic physical and biological constraints. To address this, we propose to post-train virtual cell models with reinforcement learning (RL), leveraging biologically meaningful evaluators as reward functions. We design seven rewards spanning three categories-biological function, structural validity, and morphological correctness-and optimize the state-of-the-art CellFlux model to yield CellFluxRL. CellFluxRL consistently improves over CellFlux across all rewards, with further performance boosts from test-time scaling. Overall, our results present a virtual cell modeling framework that enforces physically-based constraints through RL, advancing beyond "visually realistic" generations towards "biologically meaningful" ones.

Jeffrey Gu, Minkyu Jeon, Ambri Ma et al.

Cryo-electron microscopy (cryo-EM) is an indispensable technique for determining the 3D structures of dynamic biomolecular complexes. While typically applied to image a single molecular species, cryo-EM has the potential for structure determination of many targets simultaneously in a high-throughput fashion. However, existing methods typically focus on modeling conformational heterogeneity within a single or a few structures and are not designed to resolve compositional heterogeneity arising from mixtures of many distinct molecular species. To address this challenge, we propose CryoHype, a transformer-based hypernetwork for cryo-EM reconstruction that dynamically adjusts the weights of an implicit neural representation. Using CryoHype, we achieve state-of-the-art results on a challenging benchmark dataset containing 100 structures. We further demonstrate that CryoHype scales to the reconstruction of 1,000 distinct structures from unlabeled cryo-EM images in the fixed-pose setting.

He Li, Yuhui Zhang, Xiaohan Wang et al.

The paper demonstrate that simple adjustments of the fine-tuning recipes of multimodal large language models (MLLM) are sufficient to mitigate catastrophic forgetting. On visual question answering, we design a 2x2 experimental framework to assess model performance across in-distribution and out-of-distribution image and text inputs. Our results show that appropriate regularization, such as constraining the number of trainable parameters or adopting a low learning rate, effectively prevents forgetting when dealing with out-of-distribution images. However, we uncover a distinct form of forgetting in settings with in-distribution images and out-of-distribution text. We attribute this forgetting as task-specific overfitting and address this issue by introducing a data-hybrid training strategy that combines datasets and tasks. Finally, we demonstrate that this approach naturally extends to continual learning, outperforming existing methods with complex auxiliary mechanisms. In general, our findings challenge the prevailing assumptions by highlighting the inherent robustness of MLLMs and providing practical guidelines for adapting them while preserving their general capabilities.

Juan Manuel Zambrano Chaves, Shih-Cheng Huang, Yanbo Xu et al. · microsoft-research

The scaling laws and extraordinary performance of large foundation models motivate the development and utilization of such models in biomedicine. However, despite early promising results on some biomedical benchmarks, there are still major challenges that need to be addressed before these models can be used in real-world clinics. Frontier general-domain models such as GPT-4V still have significant performance gaps in multimodal biomedical applications. More importantly, less-acknowledged pragmatic issues, including accessibility, model cost, and tedious manual evaluation make it hard for clinicians to use state-of-the-art large models directly on private patient data. Here, we explore training open-source small multimodal models (SMMs) to bridge competency gaps for unmet clinical needs in radiology. To maximize data efficiency, we adopt a modular approach by incorporating state-of-the-art pre-trained models for image and text modalities, and focusing on training a lightweight adapter to ground each modality to the text embedding space, as exemplified by LLaVA-Med. For training, we assemble a large dataset of over 697 thousand radiology image-text pairs. For evaluation, we propose CheXprompt, a GPT-4-based metric for factuality evaluation, and demonstrate its parity with expert evaluation. For best practice, we conduct a systematic ablation study on various choices in data engineering and multimodal training. The resulting LlaVA-Rad (7B) model attains state-of-the-art results on standard radiology tasks such as report generation and cross-modal retrieval, even outperforming much larger models such as GPT-4V and Med-PaLM M (84B). The inference of LlaVA-Rad is fast and can be performed on a single V100 GPU in private settings, offering a promising state-of-the-art tool for real-world clinical applications.

Kun Yuan, Min Woo Sun, Zhen Chen et al.

There is a growing interest in developing strong biomedical vision-language models. A popular approach to achieve robust representations is to use web-scale scientific data. However, current biomedical vision-language pretraining typically compresses rich scientific figures and text into coarse figure-level pairs, discarding the fine-grained correspondences that clinicians actually rely on when zooming into local structures. To tackle this issue, we introduce Panel2Patch, a novel data pipeline that mines hierarchical structure from existing biomedical scientific literature, i.e., multi-panel, marker-heavy figures and their surrounding text, and converts them into multi-granular supervision. Given scientific figures and captions, Panel2Patch parses layouts, panels, and visual markers, then constructs hierarchical aligned vision-language pairs at the figure, panel, and patch levels, preserving local semantics instead of treating each figure as a single data sample. Built on this hierarchical corpus, we develop a granularity-aware pretraining strategy that unifies heterogeneous objectives from coarse didactic descriptions to fine region-focused phrases. By applying Panel2Patch to only a small set of the literature figures, we extract far more effective supervision than prior pipelines, enabling substantially better performance with less pretraining data.

Xiaoxian Shen, Yuhui Zhang, Sahithi Ankireddy et al.

Understanding how two radiology image sets differ is critical for generating clinical insights and for interpreting medical AI systems. We introduce RadDiff, a multimodal agentic system that performs radiologist-style comparative reasoning to describe clinically meaningful differences between paired radiology studies. RadDiff builds on a proposer-ranker framework from VisDiff, and incorporates four innovations inspired by real diagnostic workflows: (1) medical knowledge injection through domain-adapted vision-language models; (2) multimodal reasoning that integrates images with their clinical reports; (3) iterative hypothesis refinement across multiple reasoning rounds; and (4) targeted visual search that localizes and zooms in on salient regions to capture subtle findings. To evaluate RadDiff, we construct RadDiffBench, a challenging benchmark comprising 57 expert-validated radiology study pairs with ground-truth difference descriptions. On RadDiffBench, RadDiff achieves 47% accuracy, and 50% accuracy when guided by ground-truth reports, significantly outperforming the general-domain VisDiff baseline. We further demonstrate RadDiff's versatility across diverse clinical tasks, including COVID-19 phenotype comparison, racial subgroup analysis, and discovery of survival-related imaging features. Together, RadDiff and RadDiffBench provide the first method-and-benchmark foundation for systematically uncovering meaningful differences in radiological data.

Leon Liangyu Chen, Haoyu Ma, Zhipeng Fan et al.

Unified models can handle both multimodal understanding and generation within a single architecture, yet they typically operate in a single pass without iteratively refining their outputs. Many multimodal tasks, especially those involving complex spatial compositions, multiple interacting objects, or evolving instructions, require decomposing instructions, verifying intermediate results, and making iterative corrections. While test-time scaling (TTS) has demonstrated that allocating additional inference compute for iterative reasoning substantially improves language model performance, extending this paradigm to unified multimodal models remains an open challenge. We introduce UniT, a framework for multimodal chain-of-thought test-time scaling that enables a single unified model to reason, verify, and refine across multiple rounds. UniT combines agentic data synthesis, unified model training, and flexible test-time inference to elicit cognitive behaviors including verification, subgoal decomposition, and content memory. Our key findings are: (1) unified models trained on short reasoning trajectories generalize to longer inference chains at test time; (2) sequential chain-of-thought reasoning provides a more scalable and compute-efficient TTS strategy than parallel sampling; (3) training on generation and editing trajectories improves out-of-distribution visual reasoning. These results establish multimodal test-time scaling as an effective paradigm for advancing both generation and understanding in unified models.

Alejandro Lozano, Min Woo Sun, James Burgess et al. · stanford

The development of vision-language models (VLMs) is driven by large-scale and diverse multimodal datasets. However, progress toward generalist biomedical VLMs is limited by the lack of annotated, publicly accessible datasets across biology and medicine. Existing efforts are restricted to narrow domains, missing the full diversity of biomedical knowledge encoded in scientific literature. To address this gap, we introduce BIOMEDICA, a scalable, open-source framework to extract, annotate, and serialize the entirety of the PubMed Central Open Access subset into an easy-to-use, publicly accessible dataset. Our framework produces a comprehensive archive with over 24 million unique image-text pairs from over 6 million articles. Metadata and expert-guided annotations are also provided. We demonstrate the utility and accessibility of our resource by releasing BMCA-CLIP, a suite of CLIP-style models continuously pre-trained on the BIOMEDICA dataset via streaming, eliminating the need to download 27 TB of data locally. On average, our models achieve state-of-the-art performance across 40 tasks - spanning pathology, radiology, ophthalmology, dermatology, surgery, molecular biology, parasitology, and cell biology - excelling in zero-shot classification with a 6.56% average improvement (as high as 29.8% and 17.5% in dermatology and ophthalmology, respectively), and stronger image-text retrieval, all while using 10x less compute. To foster reproducibility and collaboration, we release our codebase and dataset for the broader research community.

James Burgess, Jeffrey J Nirschl, Laura Bravo-Sánchez et al. · stanford

Scientific research demands sophisticated reasoning over multimodal data, a challenge especially prevalent in biology. Despite recent advances in multimodal large language models (MLLMs) for AI-assisted research, existing multimodal reasoning benchmarks only target up to college-level difficulty, while research-level benchmarks emphasize lower-level perception, falling short of the complex multimodal reasoning needed for scientific discovery. To bridge this gap, we introduce MicroVQA, a visual-question answering (VQA) benchmark designed to assess three reasoning capabilities vital in research workflows: expert image understanding, hypothesis generation, and experiment proposal. MicroVQA consists of 1,042 multiple-choice questions (MCQs) curated by biology experts across diverse microscopy modalities, ensuring VQA samples represent real scientific practice. In constructing the benchmark, we find that standard MCQ generation methods induce language shortcuts, motivating a new two-stage pipeline: an optimized LLM prompt structures question-answer pairs into MCQs; then, an agent-based `RefineBot' updates them to remove shortcuts. Benchmarking on state-of-the-art MLLMs reveal a peak performance of 53\%; models with smaller LLMs only slightly underperform top models, suggesting that language-based reasoning is less challenging than multimodal reasoning; and tuning with scientific articles enhances performance. Expert analysis of chain-of-thought responses shows that perception errors are the most frequent, followed by knowledge errors and then overgeneralization errors. These insights highlight the challenges in multimodal scientific reasoning, showing MicroVQA is a valuable resource advancing AI-driven biomedical research. MicroVQA is available at https://huggingface.co/datasets/jmhb/microvqa, and project page at https://jmhb0.github.io/microvqa.

James Burgess, Xiaohan Wang, Yuhui Zhang et al. · stanford

How do two individuals differ when performing the same action? In this work, we introduce Video Action Differencing (VidDiff), the novel task of identifying subtle differences between videos of the same action, which has many applications, such as coaching and skill learning. To enable development on this new task, we first create VidDiffBench, a benchmark dataset containing 549 video pairs, with human annotations of 4,469 fine-grained action differences and 2,075 localization timestamps indicating where these differences occur. Our experiments demonstrate that VidDiffBench poses a significant challenge for state-of-the-art large multimodal models (LMMs), such as GPT-4o and Qwen2-VL. By analyzing failure cases of LMMs on VidDiffBench, we highlight two key challenges for this task: localizing relevant sub-actions over two videos and fine-grained frame comparison. To overcome these, we propose the VidDiff method, an agentic workflow that breaks the task into three stages: action difference proposal, keyframe localization, and frame differencing, each stage utilizing specialized foundation models. To encourage future research in this new task, we release the benchmark at https://huggingface.co/datasets/jmhb/VidDiffBench and code at http://jmhb0.github.io/viddiff.

Christopher Polzak, Alejandro Lozano, Min Woo Sun et al. · stanford

Systematic reviews (SR), in which experts summarize and analyze evidence across individual studies to provide insights on a specialized topic, are a cornerstone for evidence-based clinical decision-making, research, and policy. Given the exponential growth of scientific articles, there is growing interest in using large language models (LLMs) to automate SR generation. However, the ability of LLMs to critically assess evidence and reason across multiple documents to provide recommendations at the same proficiency as domain experts remains poorly characterized. We therefore ask: Can LLMs match the conclusions of systematic reviews written by clinical experts when given access to the same studies? To explore this question, we present MedEvidence, a benchmark pairing findings from 100 SRs with the studies they are based on. We benchmark 24 LLMs on MedEvidence, including reasoning, non-reasoning, medical specialist, and models across varying sizes (from 7B-700B). Through our systematic evaluation, we find that reasoning does not necessarily improve performance, larger models do not consistently yield greater gains, and knowledge-based fine-tuning degrades accuracy on MedEvidence. Instead, most models exhibit similar behavior: performance tends to degrade as token length increases, their responses show overconfidence, and, contrary to human experts, all models show a lack of scientific skepticism toward low-quality findings. These results suggest that more work is still required before LLMs can reliably match the observations from expert-conducted SRs, even though these systems are already deployed and being used by clinicians. We release our codebase and benchmark to the broader research community to further investigate LLM-based SR systems.

Alejandro Lozano, Min Woo Sun, James Burgess et al. · stanford

Despite the excitement behind biomedical artificial intelligence (AI), access to high-quality, diverse, and large-scale data - the foundation for modern AI systems - is still a bottleneck to unlocking its full potential. To address this gap, we introduce Biomedica, an open-source dataset derived from the PubMed Central Open Access subset, containing over 6 million scientific articles and 24 million image-text pairs, along with 27 metadata fields (including expert human annotations). To overcome the challenges of accessing our large-scale dataset, we provide scalable streaming and search APIs through a web server, facilitating seamless integration with AI systems. We demonstrate the utility of the Biomedica dataset by building embedding models, chat-style models, and retrieval-augmented chat agents. Notably, all our AI models surpass previous open systems in their respective categories, underscoring the critical role of diverse, high-quality, and large-scale biomedical data.

Josiah Aklilu, Xiaohan Wang, Serena Yeung-Levy

Precise action localization in untrimmed video is vital for fields such as professional sports and minimally invasive surgery, where the delineation of particular motions in recordings can dramatically enhance analysis. But in many cases, large scale datasets with video-label pairs for localization are unavailable, limiting the opportunity to fine-tune video-understanding models. Recent developments in large vision-language models (LVLM) address this need with impressive zero-shot capabilities in a variety of video understanding tasks. However, the adaptation of LVLMs, with their powerful visual question answering capabilities, to zero-shot localization in long-form video is still relatively unexplored. To this end, we introduce a true Zero-shot Action Localization method (ZEAL). Specifically, we leverage the built-in action knowledge of a large language model (LLM) to inflate actions into detailed descriptions of the archetypal start and end of the action. These descriptions serve as queries to LVLM for generating frame-level confidence scores which can be aggregated to produce localization outputs. The simplicity and flexibility of our method lends it amenable to more capable LVLMs as they are developed, and we demonstrate remarkable results in zero-shot action localization on a challenging benchmark, without any training. Our code is publicly available at $\href{https://github.com/josaklil-ai/zeal}{github.com/josaklil-ai/zeal}$.

Xiaohan Wang, Yuhui Zhang, Orr Zohar et al. · stanford

Long-form video understanding represents a significant challenge within computer vision, demanding a model capable of reasoning over long multi-modal sequences. Motivated by the human cognitive process for long-form video understanding, we emphasize interactive reasoning and planning over the ability to process lengthy visual inputs. We introduce a novel agent-based system, VideoAgent, that employs a large language model as a central agent to iteratively identify and compile crucial information to answer a question, with vision-language foundation models serving as tools to translate and retrieve visual information. Evaluated on the challenging EgoSchema and NExT-QA benchmarks, VideoAgent achieves 54.1% and 71.3% zero-shot accuracy with only 8.4 and 8.2 frames used on average. These results demonstrate superior effectiveness and efficiency of our method over the current state-of-the-art methods, highlighting the potential of agent-based approaches in advancing long-form video understanding.

Andong Deng, Taojiannan Yang, Shoubin Yu et al.

Large Multimodal Models (LMMs) have achieved remarkable progress across various capabilities; however, complex video reasoning in the scientific domain remains a significant and challenging frontier. Current video benchmarks predominantly target general scenarios where perception/recognition is heavily relied on, while with relatively simple reasoning tasks, leading to saturation and thus failing to effectively evaluate advanced multimodal cognitive skills. To address this critical gap, we introduce SciVideoBench, a rigorous benchmark specifically designed to assess advanced video reasoning in scientific contexts. SciVideoBench consists of 1,000 carefully crafted multiple-choice questions derived from cutting-edge scientific experimental videos spanning over 25 specialized academic subjects and verified by a semi-automatic system. Each question demands sophisticated domain-specific knowledge, precise spatiotemporal perception, and intricate logical reasoning, effectively challenging models' higher-order cognitive abilities. Our evaluation highlights significant performance deficits in state-of-the-art proprietary and open-source LMMs, including Gemini 2.5 Pro and Qwen2.5-VL, indicating substantial room for advancement in video reasoning capabilities. Detailed analyses of critical factors such as reasoning complexity and visual grounding provide valuable insights and clear direction for future developments in LMMs, driving the evolution of truly capable multimodal AI co-scientists. We hope SciVideoBench could fit the interests of the community and help to push the boundary of cutting-edge AI for border science.

Min Woo Sun, Alejandro Lozano, Javier Gamazo Tejero et al. · stanford

Embedding vision-language models (VLMs) are typically pretrained with short text windows (<77 tokens), which forces the truncation of long-format captions. Yet, the distribution of biomedical captions from large-scale open source literature reveals that a huge portion of captions far exceed 77 tokens. To this end, we investigate the impact of pretraining on long-format biomedical captions by extending the context length of text encoders in VLMs. We find that longer context (thus, enabling additional supervision provided in long-format captions) correlates with better retrieval and classification performance. Given this finding, we introduce BIOMEDICA-LongCAP, a dataset of 1M image-caption pairs enriched with context-aware descriptions from full-text articles, providing longer and additional textual supervision. Using BIOMEDICA-LongCAP, we train BMC-LongCLIP, a long-context biomedical VLM with a text encoder supporting windows of up to 512 tokens. Our model extends context capacity by 6.6x, reducing token waste from 55% to just 2.2%. On long-caption retrieval benchmarks, BMC-LongCLIP achieves up to +30% absolute gains in Recall@1 and +2% average improvements in classification, while also converging faster than short-context. Our results demonstrate that long-context modeling is a promising direction for advancing biomedical VLMs.

Shangding Gu, Xiaohan Wang, Donghao Ying et al.

Rapid advances in multimodal models demand benchmarks that rigorously evaluate understanding and reasoning in safety-critical, dynamic real-world settings. We present AccidentBench, a large-scale benchmark that combines vehicle accident scenarios with Beyond domains, safety-critical settings in air and water that emphasize spatial and temporal reasoning (e.g., navigation, orientation, multi-vehicle motion). The benchmark contains approximately 2000 videos and over 19000 human-annotated question--answer pairs spanning multiple video lengths (short/medium/long) and difficulty levels (easy/medium/hard). Tasks systematically probe core capabilities: temporal, spatial, and intent understanding and reasoning. By unifying accident-centric traffic scenes with broader safety-critical scenarios in air and water, AccidentBench offers a comprehensive, physically grounded testbed for evaluating models under real-world variability. Evaluations of state-of-the-art models (e.g., Gemini-2.5 Pro and GPT-5) show that even the strongest models achieve only about 18% accuracy on the hardest tasks and longest videos, revealing substantial gaps in real-world temporal, spatial, and intent reasoning. AccidentBench is designed to expose these critical gaps and drive the development of multimodal models that are safer, more robust, and better aligned with real-world safety-critical challenges. The code and dataset are available at: https://github.com/SafeRL-Lab/AccidentBench

Sanket Rajan Gupte, Josiah Aklilu, Jeffrey J. Nirschl et al.

Foundation vision or vision-language models are trained on large unlabeled or noisy data and learn robust representations that can achieve impressive zero- or few-shot performance on diverse tasks. Given these properties, they are a natural fit for active learning (AL), which aims to maximize labeling efficiency. However, the full potential of foundation models has not been explored in the context of AL, specifically in the low-budget regime. In this work, we evaluate how foundation models influence three critical components of effective AL, namely, 1) initial labeled pool selection, 2) ensuring diverse sampling, and 3) the trade-off between representative and uncertainty sampling. We systematically study how the robust representations of foundation models (DINOv2, OpenCLIP) challenge existing findings in active learning. Our observations inform the principled construction of a new simple and elegant AL strategy that balances uncertainty estimated via dropout with sample diversity. We extensively test our strategy on many challenging image classification benchmarks, including natural images as well as out-of-domain biomedical images that are relatively understudied in the AL literature. We also provide a highly performant and efficient implementation of modern AL strategies (including our method) at https://github.com/sanketx/AL-foundation-models.

Orr Zohar, Xiaohan Wang, Yann Dubois et al.

Despite the rapid integration of video perception capabilities into Large Multimodal Models (LMMs), the underlying mechanisms driving their video understanding remain poorly understood. Consequently, many design decisions in this domain are made without proper justification or analysis. The high computational cost of training and evaluating such models, coupled with limited open research, hinders the development of video-LMMs. To address this, we present a comprehensive study that helps uncover what effectively drives video understanding in LMMs. We begin by critically examining the primary contributors to the high computational requirements associated with video-LMM research and discover Scaling Consistency, wherein design and training decisions made on smaller models and datasets (up to a critical size) effectively transfer to larger models. Leveraging these insights, we explored many video-specific aspects of video-LMMs, including video sampling, architectures, data composition, training schedules, and more. For example, we demonstrated that fps sampling during training is vastly preferable to uniform frame sampling and which vision encoders are the best for video representation. Guided by these findings, we introduce Apollo, a state-of-the-art family of LMMs that achieve superior performance across different model sizes. Our models can perceive hour-long videos efficiently, with Apollo-3B outperforming most existing $7$B models with an impressive 55.1 on LongVideoBench. Apollo-7B is state-of-the-art compared to 7B LMMs with a 70.9 on MLVU, and 63.3 on Video-MME.

Lisa Dunlap, Yuhui Zhang, Xiaohan Wang et al. · stanford

How do two sets of images differ? Discerning set-level differences is crucial for understanding model behaviors and analyzing datasets, yet manually sifting through thousands of images is impractical. To aid in this discovery process, we explore the task of automatically describing the differences between two $\textbf{sets}$ of images, which we term Set Difference Captioning. This task takes in image sets $D_A$ and $D_B$, and outputs a description that is more often true on $D_A$ than $D_B$. We outline a two-stage approach that first proposes candidate difference descriptions from image sets and then re-ranks the candidates by checking how well they can differentiate the two sets. We introduce VisDiff, which first captions the images and prompts a language model to propose candidate descriptions, then re-ranks these descriptions using CLIP. To evaluate VisDiff, we collect VisDiffBench, a dataset with 187 paired image sets with ground truth difference descriptions. We apply VisDiff to various domains, such as comparing datasets (e.g., ImageNet vs. ImageNetV2), comparing classification models (e.g., zero-shot CLIP vs. supervised ResNet), summarizing model failure modes (supervised ResNet), characterizing differences between generative models (e.g., StableDiffusionV1 and V2), and discovering what makes images memorable. Using VisDiff, we are able to find interesting and previously unknown differences in datasets and models, demonstrating its utility in revealing nuanced insights.

Rui Li, Xiaohan Wang, Yuhui Zhang et al. · stanford

Despite significant advancements in video large multimodal models (video-LMMs), achieving effective temporal grounding in long-form videos remains a challenge for existing models. To address this limitation, we propose Temporal Preference Optimization (TPO), a novel post-training framework designed to enhance the temporal grounding capabilities of video-LMMs through preference learning. TPO adopts a self-training approach that enables models to differentiate between well-grounded and less accurate temporal responses by leveraging curated preference datasets at two granularities: localized temporal grounding, which focuses on specific video segments, and comprehensive temporal grounding, which captures extended temporal dependencies across entire video sequences. By optimizing on these preference datasets, TPO significantly enhances temporal understanding while reducing reliance on manually annotated data. Extensive experiments on three long-form video understanding benchmarks--LongVideoBench, MLVU, and Video-MME--demonstrate the effectiveness of TPO across two state-of-the-art video-LMMs. Notably, LLaVA-Video-TPO establishes itself as the leading 7B model on the Video-MME benchmark, underscoring the potential of TPO as a scalable and efficient solution for advancing temporal reasoning in long-form video understanding. Project page: https://ruili33.github.io/tpo_website.

Bingda Tang, Yuhui Zhang, Xiaohan Wang et al.

Aligning denoising generative models with human preferences or verifiable rewards remains a key challenge. While policy-gradient online reinforcement learning (RL) offers a principled post-training framework, its direct application is hindered by the intractable likelihoods of these models. Prior work therefore either optimizes an induced Markov decision process (MDP) over sampling trajectories, which is stable but inefficient, or uses likelihood surrogates based on the diffusion evidence lower bound (ELBO), which have so far underperformed on visual generation. Our key insight is that the ELBO-based approach can, in fact, be made both stable and efficient. By reducing surrogate variance and controlling gradient steps, we show that this approach can beat MDP-based methods. To this end, we introduce Variational GRPO (V-GRPO), a method that integrates ELBO-based surrogates with the Group Relative Policy Optimization (GRPO) algorithm, alongside a set of simple yet essential techniques. Our method is easy to implement, aligns with pretraining objectives, and avoids the limitations of MDP-based methods. V-GRPO achieves state-of-the-art performance in text-to-image synthesis, while delivering a $2\times$ speedup over MixGRPO and a $3\times$ speedup over DiffusionNFT.

Yuhui Zhang, Yuchang Su, Yiming Liu et al. · stanford

The rapid development of vision language models (VLMs) demands rigorous and reliable evaluation. However, current visual question answering (VQA) benchmarks often depend on open-ended questions, making accurate evaluation difficult due to the variability in natural language responses. To address this, we introduce AutoConverter, an agentic framework that automatically converts these open-ended questions into multiple-choice format, enabling objective evaluation while reducing the costly multiple-choice question creation process. Our experiments demonstrate that AutoConverter can generate correct and challenging multiple-choice questions, with VLMs demonstrating consistently similar or lower accuracy on these questions compared to human-created ones. Using AutoConverter, we construct VMCBench, a benchmark created by transforming 20 existing VQA datasets into a unified multiple-choice format, totaling 9,018 questions. We comprehensively evaluate 33 state-of-the-art VLMs on VMCBench, setting a new standard for scalable, consistent, and reproducible VLM evaluation.

Zhenzhen Weng, Jingyuan Liu, Hao Tan et al.

Reconstructing 3D humans from a single image has been extensively investigated. However, existing approaches often fall short on capturing fine geometry and appearance details, hallucinating occluded parts with plausible details, and achieving generalization across unseen and in-the-wild datasets. We present Human-LRM, a diffusion-guided feed-forward model that predicts the implicit field of a human from a single image. Leveraging the power of the state-of-the-art reconstruction model (i.e., LRM) and generative model (i.e Stable Diffusion), our method is able to capture human without any template prior, e.g., SMPL, and effectively enhance occluded parts with rich and realistic details. Our approach first uses a single-view LRM model with an enhanced geometry decoder to get the triplane NeRF representation. The novel view renderings from the triplane NeRF provide strong geometry and color prior, from which we generate photo-realistic details for the occluded parts using a diffusion model. The generated multiple views then enable reconstruction with high-quality geometry and appearance, leading to superior overall performance comparing to all existing human reconstruction methods.

Mark Endo, Xiaohan Wang, Serena Yeung-Levy

Recent works on accelerating Vision-Language Models achieve strong performance across a variety of vision-language tasks despite highly compressing visual information. In this work, we examine the popular acceleration approach of early pruning of visual tokens inside the language model. Surprisingly, we find that while strong performance is maintained across many tasks, it exhibits drastically different behavior for a subset of vision-centric tasks such as localization. Upon further investigation, we uncover a core issue with the acceleration approach where most tokens towards the top of the image are pruned away. Yet, on many benchmarks aiming to evaluate vision-centric capabilities, strong performance persists with the flawed pruning strategy, highlighting these benchmarks' limited ability to assess fine-grained visual capabilities. Based on these findings, we propose FEATHER (Fast and Effective Acceleration wiTH Ensemble cRiteria), a straightforward approach that resolves the discovered early-layer pruning issue and further enhances the preservation of relevant tokens via multistage pruning with early uniform sampling to ensure broad image coverage. With comparable computational savings, we find that FEATHER achieves more than 5x performance improvement on the vision-centric localization benchmarks compared to the original acceleration approach.

Yuhui Zhang, Yuchang Su, Chenyu Wang et al. · stanford

Building a virtual cell capable of accurately simulating cellular behaviors in silico has long been a dream in computational biology. We introduce CellFlux, an image-generative model that simulates cellular morphology changes induced by chemical and genetic perturbations using flow matching. Unlike prior methods, CellFlux models distribution-wise transformations from unperturbed to perturbed cell states, effectively distinguishing actual perturbation effects from experimental artifacts such as batch effects -- a major challenge in biological data. Evaluated on chemical (BBBC021), genetic (RxRx1), and combined perturbation (JUMP) datasets, CellFlux generates biologically meaningful cell images that faithfully capture perturbation-specific morphological changes, achieving a 35% improvement in FID scores and a 12% increase in mode-of-action prediction accuracy over existing methods. Additionally, CellFlux enables continuous interpolation between cellular states, providing a potential tool for studying perturbation dynamics. These capabilities mark a significant step toward realizing virtual cell modeling for biomedical research. Project page: https://yuhui-zh15.github.io/CellFlux/.

Binxu Li, Yuhui Zhang, Xiaohan Wang et al. · stanford

Mixed modality search -- retrieving information across a heterogeneous corpus composed of images, texts, and multimodal documents -- is an important yet underexplored real-world application. In this work, we investigate how contrastive vision-language models, such as CLIP, perform on the mixed modality search task. Our analysis reveals a critical limitation: these models exhibit a pronounced modality gap in the embedding space, where image and text embeddings form distinct clusters, leading to intra-modal ranking bias and inter-modal fusion failure. To address this issue, we propose GR-CLIP, a lightweight post-hoc calibration method that removes the modality gap in CLIP's embedding space. Evaluated on MixBench -- the first benchmark specifically designed for mixed modality search -- GR-CLIP improves NDCG@10 by up to 26 percentage points over CLIP, surpasses recent vision-language generative embedding models by 4 percentage points, while using 75x less compute.

Yuhui Zhang, Yuchang Su, Yiming Liu et al. · stanford

Negation is a fundamental linguistic phenomenon that can entirely reverse the meaning of a sentence. As vision language models (VLMs) continue to advance and are deployed in high-stakes applications, assessing their ability to comprehend negation becomes essential. To address this, we introduce NegVQA, a visual question answering (VQA) benchmark consisting of 7,379 two-choice questions covering diverse negation scenarios and image-question distributions. We construct NegVQA by leveraging large language models to generate negated versions of questions from existing VQA datasets. Evaluating 20 state-of-the-art VLMs across seven model families, we find that these models struggle significantly with negation, exhibiting a substantial performance drop compared to their responses to the original questions. Furthermore, we uncover a U-shaped scaling trend, where increasing model size initially degrades performance on NegVQA before leading to improvements. Our benchmark reveals critical gaps in VLMs' negation understanding and offers insights into future VLM development. Project page available at https://yuhui-zh15.github.io/NegVQA/.

Devanish N. Kamtam, Joseph B. Shrager, Satya Deepya Malla et al.

Background: We evaluate SAM 2 for surgical scene understanding by examining its semantic segmentation capabilities for organs/tissues both in zero-shot scenarios and after fine-tuning. Methods: We utilized five public datasets to evaluate and fine-tune SAM 2 for segmenting anatomical tissues in surgical videos/images. Fine-tuning was applied to the image encoder and mask decoder. We limited training subsets from 50 to 400 samples per class to better model real-world constraints with data acquisition. The impact of dataset size on fine-tuning performance was evaluated with weighted mean Dice coefficient (WMDC), and the results were also compared against previously reported state-of-the-art (SOTA) results. Results: SurgiSAM 2, a fine-tuned SAM 2 model, demonstrated significant improvements in segmentation performance, achieving a 17.9% relative WMDC gain compared to the baseline SAM 2. Increasing prompt points from 1 to 10 and training data scale from 50/class to 400/class enhanced performance; the best WMDC of 0.92 on the validation subset was achieved with 10 prompt points and 400 samples per class. On the test subset, this model outperformed prior SOTA methods in 24/30 (80%) of the classes with a WMDC of 0.91 using 10-point prompts. Notably, SurgiSAM 2 generalized effectively to unseen organ classes, achieving SOTA on 7/9 (77.8%) of them. Conclusion: SAM 2 achieves remarkable zero-shot and fine-tuned performance for surgical scene segmentation, surpassing prior SOTA models across several organ classes of diverse datasets. This suggests immense potential for enabling automated/semi-automated annotation pipelines, thereby decreasing the burden of annotations facilitating several surgical applications.

Jeffrey Gu, Serena Yeung-Levy

Large pre-trained models, or foundation models, have shown impressive performance when adapted to a variety of downstream tasks, often out-performing specialized models. Hypernetworks, neural networks that generate some or all of the parameters of another neural network, have become an increasingly important technique for conditioning and generalizing implicit neural representations (INRs), which represent signals or objects such as audio or 3D shapes using a neural network. However, despite the potential benefits of incorporating foundation models in hypernetwork methods, this research direction has not been investigated, likely due to the dissimilarity of the weight generation task with other visual tasks. To address this gap, we (1) show how foundation models can improve hypernetworks with Transformer-based architectures, (2) provide an empirical analysis of the benefits of foundation models for hypernetworks through the lens of the generalizable INR task, showing that leveraging foundation models improves performance, generalizability, and data efficiency across a variety of algorithms and modalities. We also provide further analysis in examining the design space of foundation model-based hypernetworks, including examining the choice of foundation models, algorithms, and the effect of scaling foundation models.

Anita Rau, Mark Endo, Josiah Aklilu et al. · deepmind, stanford

Large Vision-Language Models offer a new paradigm for AI-driven image understanding, enabling models to perform tasks without task-specific training. This flexibility holds particular promise across medicine, where expert-annotated data is scarce. Yet, VLMs' practical utility in intervention-focused domains--especially surgery, where decision-making is subjective and clinical scenarios are variable--remains uncertain. Here, we present a comprehensive analysis of 11 state-of-the-art VLMs across 17 key visual understanding tasks in surgical AI--from anatomy recognition to skill assessment--using 13 datasets spanning laparoscopic, robotic, and open procedures. In our experiments, VLMs demonstrate promising generalizability, at times outperforming supervised models when deployed outside their training setting. In-context learning, incorporating examples during testing, boosted performance up to three-fold, suggesting adaptability as a key strength. Still, tasks requiring spatial or temporal reasoning remained difficult. Beyond surgery, our findings offer insights into VLMs' potential for tackling complex and dynamic scenarios in clinical and broader real-world applications.

Zeyu Wang, Zhenzhen Weng, Serena Yeung-Levy

Conventional approaches to human mesh recovery predominantly employ a region-based strategy. This involves initially cropping out a human-centered region as a preprocessing step, with subsequent modeling focused on this zoomed-in image. While effective for single figures, this pipeline poses challenges when dealing with images featuring multiple individuals, as different people are processed separately, often leading to inaccuracies in relative positioning. Despite the advantages of adopting a whole-image-based approach to address this limitation, early efforts in this direction have fallen short in performance compared to recent region-based methods. In this work, we advocate for this under-explored area of modeling all people at once, emphasizing its potential for improved accuracy in multi-person scenarios through considering all individuals simultaneously and leveraging the overall context and interactions. We introduce a new model with a streamlined transformer-based design, featuring three critical design choices: multi-scale feature incorporation, focused attention mechanisms, and relative joint supervision. Our proposed model demonstrates a significant performance improvement, surpassing state-of-the-art region-based and whole-image-based methods on various benchmarks involving multiple individuals.

Jacob S. Leiby, Jialu Yao, Pan Lu et al.

Immunohistochemistry (IHC) provides information on protein expression in tissue sections and is commonly used to support pathology diagnosis and disease triage. While AI models for H\&E-stained slides show promise, their applicability to IHC is limited due to domain-specific variations. Here we introduce HPA10M, a dataset that contains 10,495,672 IHC images from the Human Protein Atlas with comprehensive metadata included, and encompasses 45 normal tissue types and 20 major cancer types. Based on HPA10M, we trained iSight, a multi-task learning framework for automated IHC staining assessment. iSight combines visual features from whole-slide images with tissue metadata through a token-level attention mechanism, simultaneously predicting staining intensity, location, quantity, tissue type, and malignancy status. On held-out data, iSight achieved 85.5\% accuracy for location, 76.6\% for intensity, and 75.7\% for quantity, outperforming fine-tuned foundation models (PLIP, CONCH) by 2.5--10.2\%. In addition, iSight demonstrates well-calibrated predictions with expected calibration errors of 0.0150-0.0408. Furthermore, in a user study with eight pathologists evaluating 200 images from two datasets, iSight outperformed initial pathologist assessments on the held-out HPA dataset (79\% vs 68\% for location, 70\% vs 57\% for intensity, 68\% vs 52\% for quantity). Inter-pathologist agreement also improved after AI assistance in both held-out HPA (Cohen's $κ$ increased from 0.63 to 0.70) and Stanford TMAD datasets (from 0.74 to 0.76), suggesting expert--AI co-assessment can improve IHC interpretation. This work establishes a foundation for AI systems that can improve IHC diagnostic accuracy and highlights the potential for integrating iSight into clinical workflows to enhance the consistency and reliability of IHC assessment.