NeurIPS 2026 Papers

Vision-language model (VLM) unlearning is often treated as an objective-design problem: given multimodal forget data, the goal is to optimize a loss that removes undesirable behavior while preserving retained capabilities. We argue that this view overlooks an equally important data-centric question: how should each forget example be represented before optimization? Existing methods typically keep image-conditioned failures in their original image-text form, implicitly assuming that pixels are the right route for unlearning whenever pixels appear in the input. We challenge this assumption by constructing modality-decomposed versions of VLM benchmarks, enabling controlled comparisons among image-only, text-only, and multimodal representations of the same examples. Our modality-mixing experiments show that, for many examples, textual renderings carry the actionable forget signal more directly than the image alone, while purely text-only unlearning can still weaken visual grounding when the target behavior depends on visual evidence. Motivated by this trade-off, we introduce ModRoute, a gradient-guided per-sample modality routing strategy that selectively converts high visual-pressure examples to text-only form while keeping the remaining examples multimodal. Notably, on VLGuard with LLaVA-1.5-7B, ModRoute has the strongest composite unlearning--utility score of $0.847$ at $\gamma=0.5$, a $2.5$% improvement over the best random-switching score and $\geq46$% better than pure multimodal or text unlearning. Overall, our results show that effective VLM unlearning depends not only on the objective, but also on the composition and per-sample modality routing of the unlearning data.

Sliding-window human activity recognition (HAR) splits continuous sensor streams into fixed-length, overlapping windows and trains a classifier to assign one activity label to each window. Window length and stride are chosen from sampling rate, latency or computation constraints, and assumptions about typical motion duration, but they need not align with motion cycles or activity changes. Some windows therefore contain only part of a repeated motion, fragments shared by related activities, or mixed evidence near a transition. Yet each window still receives a single label, creating a granularity mismatch between activity-level supervision and local motion evidence. We propose PSDNet (Phase-Script Deliberation Network), which separates local phase interpretation from final activity classification. PSDNet encodes the current window into latent phase primitives, combines them with recent history, and produces an initial classification with an uncertainty score. When uncertainty remains high, a deliberation module compares the current phase evidence with learned class-specific phase scripts, i.e., compact templates of short-term phase patterns, and updates classification scores over a small candidate set. An auxiliary boundary objective encourages sensitivity to possible activity changes without serving as a hard routing rule. This design concentrates extra computation where direct classification is least reliable while keeping easy cases efficient. Experiments on eight public HAR benchmarks show that PSDNet consistently improves accuracy and weighted F1 over strong CNN- and sequence-based baselines. Additional analyses on ambiguous windows and similar-activity confusion pairs support phase-aware selective deliberation for windows whose local evidence is insufficient for reliable classification.

Vision-language models (VLMs) have achieved significant progress in scene understanding and spatial reasoning. However, existing evaluation benchmarks predominantly assume that all visible content corresponds directly to physical entities in the scene, leaving model performance in reflective scenarios largely unexamined. Reflective surfaces preserve object appearance while altering the mapping between appearance and physical location, causing VLMs to produce systematic reasoning errors in such scenes. In this work, we introduce (i) a grounded RGB-D dataset for mirror-aware spatial reasoning, and (ii) the first comprehensive and systematic evaluation of state-of-the-art VLMs in reflective scenarios. Specifically, we collect 1,011 RGB-D images from 525 real-world scenes, spanning diverse indoor and outdoor environments and six reflective surface types, under varying lighting conditions across both daytime and nighttime settings. All images are annotated with reflective surface region masks and fine-grained, polygon-level instance annotations for both directly observed and mirror-reflected objects, with depth maps serving as both geometric ground truth for annotation and optional model input. We curate a mirror-induced reasoning ambiguity benchmark, named \textbf{MIRA-Bench} (Mirror-Induced Reasoning Ambiguity Benchmark). MIRA-Bench comprises 2,955 unique QA pairs across eight sub-tasks organized into three cognitive levels: \mbox{Reflection-aware} Perception, Spatial Reasoning, and \mbox{Scene-level} \mbox{Decision-making}, with five object-referencing protocols yielding 14,140 QA samples. We evaluate state-of-the-art VLMs on MIRA-Bench and find that these models consistently underperform across tasks, suggesting that current VLMs do not yet understand mirrors. Common failure patterns include misclassifying whether objects are directly observed or seen through reflections, failing to identify that reflected and directly observed instances correspond to the same physical object, producing erroneous cross-space spatial inferences, and tending to adopt conservative default strategies in action-oriented decisions instead of reasoning about the true scene layout. We will release the dataset and evaluation code to support future research.

Multimodal Large Language Models (MLLMs) still struggle with fine-grained visual understanding, where answers often depend on small but decisive evidence in the full image. We observe a regional-to-global perception gap: the same MLLM answers fine-grained questions more accurately when conditioned on evidence-centered crops than on the corresponding full images, suggesting that many failures stem from difficulty to focus on relevant evidence rather than insufficient local recognition ability. Motivated by this observation, we propose Vision-OPD (Vision On-Policy Distillation), a regional-to-global self-distillation framework that transfers the model's own privileged regional perception to its full-image policy. Vision-OPD instantiates two conditional policies from the same MLLM: a crop-conditioned teacher and a full-image-conditioned student. The student generates on-policy rollouts, and Vision-OPD minimizes token-level divergence between the teacher and student next-token distributions along these rollouts. This enables the model to internalize the benefit of visual zooming without external teacher models, ground-truth labels, reward verifiers, or inference-time tool use. Experiments on multiple fine-grained visual understanding benchmarks show that Vision-OPD models achieve competitive or superior performance against much larger open-source, closed-source, and "Thinking-with-Images" agentic models.

Vision-language grounding connects language to visual content, yet most existing formulations reduce grounding to a unidirectional localization problem: given a prespecified text phrase or category name, identify the corresponding image region. This setup assumes that the relevant linguistic unit is already known, overlooking a more basic challenge in grounded communication: determining which parts of the text are visually referential and how they correspond to entities in the image. We formulate grounding as bidirectional concept correspondence over an image-text pair. Given an image and its paired text, the goal is to recover all correspondences between visually referential text spans and instance-level image segments, without assuming that the relevant text spans are provided. This formulation unifies common grounding tasks, including phrase grounding, referring expression grounding, and open-vocabulary detection, by treating text segmentation, image segmentation, and cross-modal alignment as a single correspondence prediction problem. To address this task, we introduce our model, a grounding model built on top of a pretrained vision-language model. It uses learnable bridge tokens to represent candidate image-text correspondences and predicts, for each token, a text mask, an image mask, and a correspondence presence score. To train and evaluate this task, we convert diverse grounding and segmentation datasets into a unified correspondence format. Experiments show that our model consistently outperforms baselines, improving correspondence F1 by 59% on the long-caption dataset and by 73% on zero-shot LVIS, where the large category list serves as the text input.

Figure-ground organization in the human visual system relies on several shape-based cues, including surroundedness, convexity, and symmetry. While these cues have been extensively studied using abstract stimuli, little is known about how they operate under natural conditions or how they arise from the statistics of natural scenes. Deep neural networks offer a promising path forward: a model that relies on the same figure-ground cues as humans would provide tractable experimental access to the underlying mechanisms. In this study, we evaluate shape-based figure-ground organization in Vision Transformers (ViTs), for which prior work has demonstrated the emergence of object-based grouping. We test 25 ViTs spanning supervised and self-supervised training objectives, by fitting linear probes to predict figure-ground assignment from intermediate patch representations using both natural images and controlled artificial stimuli that isolate individual cues. Our results show that ViTs robustly encode surroundedness and convexity, and that probes trained on natural images generalize zero-shot to artificial stimuli across several models. For symmetry we observe mixed results: the cue is encoded for uniformly colored but not for textured regions. Taken together, our findings demonstrate that Gestalt-like figure-ground cues can be learned from natural scene statistics and position ViTs as a compelling model system for studying the computational mechanisms of perceptual organization. All code and data will be published.

Large vision-language models (VLMs), which process both visual inputs and text queries, incur high inference costs due to the large number of visual tokens, making visual token pruning a natural approach for improving efficiency. Recent work has leveraged query information to guide token selection, often through heuristics derived from attention patterns or token dynamics. However, these approaches do not explicitly exploit the underlying structure of query-vision interactions. We address this gap by observing that the query-vision interaction matrix, naturally induced by the query and key projections already present in transformer attention, is effectively low-rank, with a small number of dominant latent interaction modes capturing query-relevant semantics. Motivated by this observation, we propose Query-Vision Decomposition (QViD), a novel training-free query-aware visual token pruning method that exploits this structure. Extensive experiments on both image and video understanding benchmarks show that QViD consistently improves performance under matched token budgets, with the largest gains appearing under aggressive compression regimes.

Vision Transformers (ViTs) based vision foundation models (VFMs) have achieved remarkable performance across diverse vision tasks, but suffer from quadratic complexity that limits scalability to long sequences. Existing linear attention approaches for ViTs are typically trained from scratch, requiring substantial computational resources, while linearization-based methods developed for large language model decoders do not transfer well to ViTs. To address these challenges, we propose ViT-AdaLA, a novel framework for effectively adapting and transferring prior knowledge from VFMs to linear attention ViTs. ViT-AdaLA consists of three stages: attention alignment, feature alignment, and supervised fine-tuning. In the attention alignment stage, we align vanilla linear attention with the original softmax-based attention in each block to approximate the behavior of softmax attention. However, residual approximation errors inevitably accumulate across layers. We mitigate this by fine-tuning the linearized ViT to align its final-layer features with a frozen softmax VFM teacher. Finally, the adapted prior knowledge is transferred to downstream tasks through supervised fine-tuning. Extensive experiments on classification and segmentation tasks demonstrate the effectiveness and generality of ViT-AdaLA over various state-of-the-art linear attention counterpart.

LiDAR localization is a fundamental task in robotics and computer vision, aiming to estimate the global pose of point clouds. Although Scene Coordinate Regression (SCR) has demonstrated state-of-the-art performance in this field, standard SCR methods employ a monolithic network for uniform optimization across diverse scenes, which inevitably suffers from capacity interference in highly degraded scenes. Recent research on Vision-Language Models (VLMs) indicates that they can acquire rich scene understanding priors, providing the adaptive learning capabilities currently missing in localization networks. In this paper, we propose ViLo, the first framework to integrate VLM priors into SCR to fundamentally enhance localization robustness. Specifically, we design a temporally consistent key-frame querying mechanism to extract open-vocabulary degeneracy priors. We then construct a dynamic prior-guided Mixture-of-Experts (MoE) model for adaptive feature learning tailored to varying scenes. Finally, to avoid additional computational overhead during inference, ViLo internalizes the VLM's scene understanding into the 3D backbone via cross-modal distillation, enabling a lightweight, LiDAR-only deployment. Extensive experiments on the Oxford RobotCar and NCLT datasets demonstrate that ViLo significantly outperforms existing state-of-the-art methods, reducing positional errors by impressive margins of 20% and and 51%, respectively.

Vision-Language-Action (VLA) models have emerged as powerful foundations for robotic manipulation, but their reliance on fixed camera configurations during training makes them brittle to changes in camera count or pose during deployment. To overcome these limitations, we propose VersaCamVLA, a camera-configurable framework that decouples camera-set representation from action learning. VersaCamVLA learns a unified scene-token interface that maps an arbitrary, variable set of posed RGB views into fixed-size latent scene tokens. This is achieved via multi-signal target-view prediction and Wrist-Augmented Pose Sampling (WAPS), which leverages natural wrist-camera motion for free pose diversity. At deployment, a lightweight spatial encoder injects these compact scene tokens into a pretrained base VLA as a supplementary visual condition, requiring no explicit 3D sensing or novel-view rendering. Experiments on RoboTwin, LIBERO, and a real-robot platform demonstrate that VersaCamVLA consistently outperforms prior VLA methods and direct multi-view baselines, maintaining robust performance across varying camera counts and unseen camera poses.

Recent advancements in Multimodal Large Language Models (MLLMs) have enabled complex reasoning. However, existing remote sensing (RS) benchmarks remain heavily biased toward perception tasks, such as object recognition and scene classification. This limitation hinders the development of MLLMs for cognitively demanding RS applications. To address this, we propose a Vision Language ReaSoning Benchmark (VLRS-Bench), which is the first benchmark exclusively dedicated to complex RS reasoning. Structured across the three core dimensions of Cognition, Decision, and Prediction, VLRS-Bench comprises 2,000 question-answer pairs with an average question length of 130.19 words, spanning 14 tasks and up to eight temporal phases. VLRS-Bench is constructed via a specialized pipeline that integrates RS-specific priors and expert knowledge to ensure geospatial realism and reasoning complexity. Experimental results reveal significant bottlenecks in existing state-of-the-art MLLMs, providing critical insights for advancing multimodal reasoning within the remote sensing community.

Vision Language Models (VLMs) enable a unified model to solve various vision tasks through prompting. They have shown promising performance in semantic understanding. However, 3D understanding still largely relies on expert vision models with complex task-specific designs. The key argument this work wants to make is that VLMs are native 3D learners. Our in-depth large scale study shows that 1) focal length unification, 2) text-based pixel reference and 3) data mixture and scaling, are all you need for effective 3D learning. Model architecture changes, large models, heavy data augmentations, and complex losses including the regression formulation, many of which form the foundation of expert vision models, are actually not necessary conditions. As a result, we propose VLM$^3$, a scalable method with the simplest design that enables standard VLMs to master diverse 3D tasks. VLM$^3$ not only advances the VLM depth estimation accuracy by a large margin (0.84 $\rightarrow$ 0.9), but also enables diverse 3D tasks such as pixel correspondence, camera pose estimation and object-level 3D understanding, matching expert vision model accuracy while maintaining standard architectures and text-based training. Code will be released to the community.

Vision-Language-Action (VLA) models have shown strong promise for general-purpose robotic manipulation, but their real-world evaluation remains limited by a lack of accessible, reproducible, and consistent benchmarks. Simulation benchmarks fail to capture real-world complexity, while existing real-world benchmarks often require expensive hardware, centralized evaluation, or are limited in task diversity. We introduce *VLA-REPLICA*, a low-cost, easily reproducible real-world benchmark for evaluating VLA models. Built from off-the-shelf components, our system can be quickly assembled and replicated across laboratories, providing a consistent environment for policy evaluation anywhere in the world. *VLA-REPLICA* includes a diverse suite of manipulation tasks and a small-scale demonstration dataset for target-domain adaptation, with real-world evaluation protocols for both in-distribution and out-of-distribution settings. Experiments with imitation learning and state-of-the-art VLA models reveal model strengths and limitations, while consistent results across independently constructed setups demonstrate the reproducibility of our benchmark.

Flow-based vision-language-action (VLA) policies provide a compelling recipe for robot control: they inherit broad behavioral priors from large-scale imitation and generate smooth, temporally extended continuous action chunks. The same properties, however, make offline post-training delicate. Mixed offline robot datasets contain expert behavior as well as partial completions and failures, requiring improvement without drifting far from the data support. An offline critic can in principle provide useful directions, but actor-side critic optimization may suffer from extrapolation and push high-dimensional action chunks off-support. In contrast, advantage-conditioned or reweighting-based methods are more stable, but tend to be conservative and restricted to actions already present in the dataset. We propose VISTA, a post-training framework that turns value estimates into support-anchored distillation targets for pretrained flow policies. VISTA queries a critic only at dataset actions, computes a normalized action-gradient direction, and uses this bounded direction to shift clean-action endpoints and velocity targets. A frozen flow teacher preserves the pretrained action prior, while a behavior-cloning anchor keeps the adapted targets close to the data distribution. The shaped supervision is distilled into a time-indexed action denoising (TAD) student, which predicts time-indexed clean-action estimates with a single network evaluation and obtains the executed action by an analytic final readout. We treat the induced target shift as conservative local guidance rather than a global policy-improvement guarantee. Across five BEHAVIOR-1K simulation tasks and five real-world bimanual manipulation tasks, VISTA achieves the best average success rate in our evaluation against behavior cloning (BC), advantage-conditioned BC, IDQL, and flow Q-learning, while reducing action-expert computation by $15\times$ (and end-to-end policy inference by $1.8\times$) relative to the $20$-step flow teacher.

Pathology vision-language models are advancing rapidly, yet existing benchmarks remain focused on human tissue, particularly oncology, leaving non-human pathology largely unaddressed. This gap is especially important in toxicologic pathology, where microscopic tissue examination of laboratory animals is a core component of preclinical drug safety assessment. To address it, we introduce VIPER, the first expert-curated benchmark for vision-language model evaluation in toxicologic pathology. VIPER contains 1251 questions associated with 419 H&E-stained rat histology images across 9 organs, covering multiple-choice, KPrim, and free-text formats. All questions were curated and validated by board-certified veterinary pathologists. In total, we benchmarked 16 models, including two newly introduced veterinary-pathology models, seven human pathology-specialized models, and seven general-purpose frontier models. The results identify a substantial domain gap between veterinary and human pathology, expose the risk of over-diagnosis of normal tissue in frontier models, and show that domain-specific training remains critical for visually grounded predictions. VIPER data and evaluation code are available at https://github.com/mahmoodlab/viper.

Scalable Vector Graphics (SVG) are essential for technical illustration and digital design, offering resolution independence and semantic editability. In practice, original vector files are frequently lost, leaving only rasterized versions (e.g., PNG, JPEG) that resist modification, while manual reconstruction is prohibitively expensive. Progress on automating raster-to-SVG conversion has been bottlenecked by two gaps: existing SVG datasets are dominated by icons and decorative graphics that lack the complexity of professional diagrams, and existing benchmarks rely on pixel- or embedding-level similarity that fails to capture structural correctness (e.g., broken connectivity, misplaced arrows). We close both gaps with paired contributions targeting diagram-centric figures (e.g., model architectures, flowcharts, schematics). For training, we introduce VFIG-Data, the largest figure-to-SVG dataset of its kind at 66K pairs, combining real paper figures converted via a describe-and-generate pipeline with programmatic diagrams that supply noise-free supervision over arrow styles, fonts, and geometry. For evaluation, we introduce VFIG-Bench, a structure-aware evaluation suite, paired with VFIG-Bench-OOD, an out-of-distribution set of figures manually curated from highly cited arXiv papers. Beyond pixel and embedding similarity, our protocol reports rubric-based VLM-Judge scores and Elo ratings from pairwise human preference evaluation. Built on these contributions, VFIG is a VLM family trained with a simple-to-complex SFT curriculum followed by RL with rendering-aware rewards. VFIG achieves state-of-the-art open-source performance, outperforming the best open-source VLM baseline by over 30\%, and matches Claude Sonnet 4.6 on VFIG-BENCH: Gemini-Judge 78.2% vs. 76.7% and GPT-Judge 87.5% vs. 87.4%. It remains slightly behind the strongest proprietary models GPT-5.2 and Gemini-3.

Large-scale video–text contrastive learning has become the de facto standard for video understanding, yet mainstream methods almost universally adopt the linear patch embedding of vision Transformers (ViT) or evenly sliced tubelets as the front end. We revisit this front end from the biological principles of the primary visual cortex (area V1) and introduce V1DynamicVisionSystem (V1-DVS), a fully end-to-end trainable, bio-plausible video embedding framework. Its main contributions are: (i) a three-pathway parallel architecture that maps onto layer 4C orientation selectivity, layer 4B motion direction selectivity, and the cytochrome-oxidase Blob color-opponent system; (ii) an explicit shape–color decoupling, with isotropic Blob processing running in parallel with anisotropic Gabor orientation processing; (iii) long-range spatiotemporal convolution that captures inter-frame optical flow through 3D filters and the Adelson–Bergen motion-energy model rather than naive frame stacking; and (iv) bidirectional cross-pathway attention, allowing motion and color features to mutually modulate, as in V1→V4 inter-areal feedback. Pretrained on CC3M, WebVid-10M, ActivityNet and Ego4D and evaluated under a fully self-supervised objective(SimCLR + temporal-order verification + masked-token prediction), V1-DVS outperforms four strong baselines—linear patch embedding, mean temporal pooling, tubelet embedding, and spatiotemporal latent encoding—on six metrics spanning retrieval and embedding-space geometry, validating the practical value of biologically grounded structure for video embedding.

Multimodal large language models are increasingly expected to perform \emph{thinking with images}, yet existing visual latent reasoning methods still rely on explicit textual chain-of-thought interleaved with visual latent tokens. This interleaved design limits efficiency and keeps reasoning fragmented across separate text and vision channels. We propose **UniVLR**, a unified visual latent reasoning framework that treats textual reasoning and auxiliary visual evidence as a shared visual workspace. Instead of preserving text CoT as an independent inference-time path, UniVLR renders reasoning traces together with auxiliary images and learns to compress this unified representation into compact visual latent tokens. At inference time, the model reasons only through visual latents and directly decodes the final answer, avoiding both external tool calls and verbose text reasoning. Experiments on real-world perception and visual reasoning tasks show that UniVLR outperforms prior visual latent reasoning methods while using substantially fewer generated reasoning tokens, suggesting a more unified and efficient paradigm for visual thinking in MLLMs. Our code is available at: https://anonymous.4open.science/r/UniVLR-1B0C/.

Decades of cognitive science establish that humans navigate environments by forming cognitive maps, defined as allocentric and topology-preserving representations of 3D space. While modern Vision-Language Models (VLMs) demonstrate emergent spatial reasoning from 2D egocentric inputs, it remains unclear whether they construct an analogous 3D internal representation. In this paper, we demonstrate that current VLMs do possess a latent topological map of 3D scenes, but it is heavily overshadowed by non-geometric visual semantics, such as color and shape. By isolating this spatial subspace through cross-scene linear feature extraction, we extract a clean spatial subspace that causally controls the model's spatial outputs. We mathematically shape this latent representation and prove its correspondence to the Laplacian eigenmaps of the scene's 3D Gaussian-kernel graph, converging to the physical 3D space in the continuous limit. Motivated by this geometric identification, we further introduce a mathematically principled latent regularization method for VLMs, based on Dirichlet energy. Applying this single-term regularizer to a minimal 500-step supervised VLM fine-tuning (SFT) on simple synthetic data yields significant improvements on real-world spatial benchmarks, outperforming standard SFT and competitive baselines by up to 12.1\% in spatial tasks involving scene topology understanding.