NeurIPS 2026 论文列表

Video Reasoning Segmentation (VRS) demands both complex semantic reasoning and robust spatiotemporal mask propagation. Existing paradigms typically compress reasoning into an isolated, opaque token (e.g., \texttt{[SEG]}) via expensive fine-tuning, or rely on training-free prompt-then-track pipelines that inevitably suffer from semantic degradation and identity switches when encountering hard distractors. To fundamentally address these vulnerabilities, we propose Counterfactual Logic and Explicit Anchoring Reasoning (\textbf{CLEAR}) framework, a training-free VRS framework inspired by human dual-system cognition. Specifically, we introduce a Causal-Counterfactual Inference mechanism that transforms error-prone coordinate regression into a tractable discrete instance selection task, establishing highly reliable initial anchors via rigorous bidirectional reasoning. During mask propagation, we design an adaptive mode-switching mechanism bridged by a Heterogeneous Event Monitor. The fast-thinking mode utilizes a purely visual engine for efficient temporal propagation, while the Monitor continuously evaluates its reliability across multiple dimensions. Upon detecting propagation anomalies, the slow-thinking mode intervenes and explicitly replaces \texttt{[SEG]} tokens with a structured Contrastive Concept Dictionary for closed-loop identity recovery..Extensive experiments demonstrate that CLEAR significantly outperforms existing training-free methods and achieves performance comparable to state-of-the-art training-based approaches on comprehensive referring and reasoning VOS benchmarks. Code will be available after acceptance.

Training LLMs to think and reason for longer has become a key ingredient in building state-of-the-art models that can solve complex problems previously out of reach. Recent efforts pursue this in different ways, such as RL fine-tuning to elicit long CoT or scaling latent reasoning through architectural recurrence. This makes reasoning length an important scaling knob. In this work, we identify a novel phenomenon (both theoretically and experimentally): under outcome-only supervision, out-of-distribution (OOD) performance can continue improving as training-time reasoning length (e.g., the token budget in RL, or the loop count in looped Transformers) increases, even after in-distribution (ID) performance has saturated. This suggests that robustness may require a larger budget than ID validation alone would indicate. We provide theoretical explanations via two mechanisms: (i) self-iteration can induce a stronger inductive bias in the hypothesis class, reshaping ID-optimal solutions in ways that improve OOD generalization; and (ii) when shortcut solutions that work for ID samples but not for OOD samples persist in the hypothesis class, regularization can reduce the learned solution's reliance on these shortcuts as the number of self-iterations increases. We complement the theory with empirical evidence from two realizations of scaling training-time reasoning length: increasing the number of loops in looped Transformers on a synthetic task, and increasing token budgets during RL fine-tuning of LLMs on mathematical reasoning, coding, and cross-representation tasks.

Understanding long-form video remains a fundamental challenge for multimodal large language models (MLLMs). Sparse frame sampling fails to capture fine-grained visual details, while dense sampling quickly exceeds context length limits. Retrieval-augmented generation (RAG) offers a promising middle ground by selectively retrieving relevant video segments for grounded generation, yet its effectiveness critically depends on the quality of the video segments used as retrieval units. In this paper, we investigate RAG for movie understanding, which demands story-level reasoning over characters, events, and narrative arcs spanning hours of content. Scene segmentation, a long-studied problem that partitions movies into semantically coherent units, is a natural candidate for defining such retrieval units. We reexamine whether existing methods actually serve this role through comprehensive evaluation on downstream movie understanding tasks, and find that they consistently fail to outperform naive uniform temporal chunking. Our audit of the most standard scene segmentation benchmarks reveals why: current annotations prioritize visually salient transitions over narrative event structure. Motivated by this mismatch, we introduce NarraScene, a narrative-centric scene segmentation dataset annotated with a three-level cognitive taxonomy spanning physical, character, and narrative change, where every valid boundary requires a narrative-level shift. When used as retrieval units, these narrative-grounded segments outperform uniform chunking on downstream movie understanding tasks, suggesting that the central challenge for scene segmentation in movie RAG is not detecting boundaries, but identifying the narrative event units that matter for movie understanding.

Cell-level dense prediction is central to computational pathology, but remains challenging due to fine-grained histological structures, strong domain shifts, and costly dense annotations. Existing ViT-based pathology foundation models rely on patch tokenization, which can disrupt spatial continuity and weaken local morphological details needed for cell-level prediction. To address this, we propose Masked-Diffusion Convolutional Foundation Models, termed ConvNeXt Masked-Diffusion (CMD), a self-supervised convolutional generative pretraining framework for dense pathology representation learning. CMD uses a fully convolutional ConvNeXt-UNet backbone, performs masked-diffusion pretraining in pixel space, and incorporates frozen pathology foundation model features through adaptive normalization. Experimental results demonstrate that CMD consistently outperforms existing ViT-based pathology foundation models and even surpasses state-of-the-art end-to-end segmentation methods while fine-tuning only a small number of task-specific parameters across multiple pathology dense prediction tasks. The advantage is particularly pronounced under limited annotation settings, where CMD exhibits stronger robustness and generalization ability. Our findings suggest that purely convolutional architectures can also serve as competitive pathology foundation models for cell-level dense prediction, achieving leading performance within the current ViT-dominated paradigm and providing a scalable, high-performance solution that better preserves histological structural priors for fine-grained pathology understanding. Project page: https://anonymous.4open.science/r/ConvNeXt-Masked-Diffusion-5BC6.

Existing hallucination detectors for retrieval-augmented generation (RAG), whether based on model outputs (e.g., likelihood) or internal representations, typically apply a uniform detection strategy across responses. However, we find that different detection signals are informative in different response regimes, making a uniform strategy insufficient for both short- and long-form generation. In short-form tasks such as factoid QA, hallucination typically appears as selecting an incorrect final answer among multiple context-relevant candidates. In this regime, model-output-based signals that discriminate among plausible answers are particularly important. We therefore propose a bidirectional likelihood measure that evaluates logical consistency based on the retrieved evidence and a reasoning step generated after the response. In contrast, long-form descriptive responses are more prone to gradual drift: as generation proceeds, the model increasingly relies on its own prior generations and parametric knowledge, while unsupported continuations may remain locally plausible, making output-based cues less informative. For this regime, we measure context--knowledge conflict by tracking the directional alignment of context-grounding versus internal-knowledge contributions in the hidden states. Building on these regime-specific signals, we introduce ARGUS, an adaptive hallucination detector that emphasizes the more effective signal according to the response regime. Experiments on multiple RAG benchmarks show that our method achieves strong performance across both short- and long-form settings.

Existing underwater image enhancement (UIE) methods predominantly rely on unidirectional mapping, which lacks often results in over-enhancement or under-enhancement, without user-friendly controllability. To address this, we propose a paradigm of unsupervised trajectory learning for Omnidirectional Controllable UIE that treats restoration as a omnidirectional dynamical system, namely OC-UIE. The main purpose of our work is the realization of a consistent omnidirectional mapping across the representation space, which allows the model to master complex underwater dynamics beyond traditional unidirectional constraints. To implement this, we propose an Omnidirectional Training that optimizes over arbitrary source-target positions sampled from the data distribution. We decompose this purpose into two synergistic stages: first, Representation Trajectory Flattening is employed to organize intricate degradations into a unified adaptation axis; second, Trajectory Omnidirectional Integration is introduced to model the enhancement as a path integral over the resulting manifold. By optimizing only the relative shifts, OC-UIE isolates structural preservation from fluid degradation dynamics. This approach provides a controllable space for intensity-parameterized enhancement with state-of-the-art performance.

As large language models (LLMs) become more capable and agentic, the requirement for trust in their outputs grows significantly, yet at the same time concerns have been mounting that models may learn to lie in pursuit of their goals. To address these concerns, a body of work has emerged around the notion of "honesty" in LLMs, along with interventions aimed at mitigating deceptive behaviors. However, some benchmarks claiming to measure honesty in fact simply measure accuracy—the correctness of a model's beliefs—in disguise. Moreover, no benchmarks currently exist for directly measuring whether language models overtly lie. In this work, we introduce a realistic, large-scale human-curated dataset for directly measuring lies of commission, allowing us to disentangle accuracy from honesty. Across a diverse set of LLMs, we find that while larger models obtain higher accuracy on our benchmark, they do not become more honest. Surprisingly, most frontier LLMs obtain high scores on truthfulness benchmarks yet exhibit a substantial propensity to lie under pressure, resulting in low scores on our benchmark. We find that simple methods, such as representation engineering interventions, can improve honesty. These results underscore the growing need for robust evaluations and effective interventions to ensure LLMs remain trustworthy.

We argue that cooperative multi-agent reinforcement learning (MARL) should treat test-time coordination as part of the problem definition rather than a deployment artifact. Our focus is on long-horizon, centrally monitored cooperative systems such as warehouse robotics, automated storage-and-retrieval systems, and cloud-resource scheduling, where teams share a reward, operate under shifting objectives, and already run with global telemetry at deployment. The dominant paradigm in this regime, centralized training with decentralized execution (CTDE), uses global information only during training and provides no coordination channel at execution. This is sufficient when agents are weakly coupled or the environment is stationary, but it fails when the team must realign after conditions change. The limitation is structural rather than algorithmic, and more training cannot add a channel that does not exist at execution time. We call this the recoverability gap. To address it, we name the underexplored design space of intermittent test-time coordination as centralized training with mixed execution (CTME), in which a coordinator observes global state and emits high-level signals while agents act locally. CTME is not a particular learning algorithm, but a cooperative MARL problem setting that makes test-time information structure explicit. It specifies team-level information, high-level signals, and update frequency after training, with CTDE recovered as a limit case. We do not argue that CTDE is obsolete. We argue that removing the test-time channel is a modeling choice, not a necessity, and the wrong default when global telemetry is available and teams must recover from drift.

Cognitive primitives such as planning, exploration, and backtracking are widely regarded as core to competent web agents. Yet existing studies are largely retrospective: agents are evaluated on domain-based tasks, and trajectories are inspected post hoc for evidence of these behaviors. Because such tasks are not designed to require specific primitives, observed behaviors do not reliably reflect true capability demands. This raises a key question: Can we design realistic tasks where success provably depends on a chosen cognitive primitive? WEBSTRESS addresses this by casting capability evaluation as a controlled comparison. For each task, we construct a paired variant in which a target capability becomes necessary (e.g., transforming a direct path into an obstructed one requiring exploration). The performance gap between the pair measures the agent’s reliance on that capability. Across seven environments and seven capabilities, this yields 519 paired tasks. Evaluating six frontier agents reveals consistent weaknesses: all suffer 18-28\% drops on capability-demanding variants, with 75\% of failures being belief failures, where agents incorrectly declare success. In contrast, humans exhibit only a 5.7\% performance drop under the same paired evaluation. These results highlight that, despite strong overall performance, substantial gaps remain between human and agent cognitive behavior, and that our approach provides a principled way to measure them.

Long-horizon agents increasingly operate across many steps, tools, and observations. In this setting, the relevant oversight question is not only whether each action is locally valid, but whether the evolving trajectory still corresponds to the task the user authorized. Drift can accumulate quietly: an agent may call the right tool with plausible arguments at every step, while its prefix moves toward a broader role, an adjacent objective, or evidence the user never supplied. Existing monitors mostly check local compliance, deliver final-trace verdicts, or score generic risk; they do not directly estimate this prefix-level relation. We introduce ontological trust, a task-conditioned property of trajectory prefixes, and instantiate it as RGE, an online monitor that decomposes trust along Role, Goal, and Evidence. RGE uses LLMs only to derive structured representations. State updates, projections, and intervention decisions are deterministic, so the output is a replayable and auditable trust trajectory rather than a single end-to-end judge verdict. We construct a cross-domain trajectory corpus from OSWorld, FinanceBench, and EICU-AC, covering benign executions, prefix-paired drift, and pseudo-consistency failures. On this corpus, RGE outperforms adapted rule-, judge-, and shield-style baselines on prefix-paired drift detection. With the two larger estimator models, it exceeds 93\% Drift F1 on every benchmark while keeping benign coverage at or above 95.8\%. Pseudo-consistency is harder: detection depends on whether task completion is externally visible, a structural limit we characterize empirically.

Offline goal-conditioned reinforcement learning (GCRL) is typically benchmarked by the best tuned success rate of each method. This score measures attainable performance, but it does not reveal how reliably a learned goal-conditioned signal can be extracted into a policy: a method could succeed across many value-learning and extraction settings, or only at a narrow, hard-to-find configuration. We study this gap across four methods, GCIQL, GCIVL, QRL, and CRL, under a shared advantage-weighted regression (AWR) extractor. For each method, we construct trainability landscapes over the optimizer learning rate, which affects value learning and actor optimization, and AWR temperature, which controls how selectively the actor imitates high-advantage transitions. Across AntMaze, Cube, and Scene, we observe distinct regimes: high-scoring methods may be broadly accessible or brittle, while broad relative basins may still sit below low absolute ceilings. To interpret these differences, we pair landscapes with post-hoc diagnostics of future-vs-random goal discrimination and AWR weight concentration. % whether the learned advantage function separates matched future goals from random goals, and whether AWR spreads weight across many behavior-supported transitions or concentrates on a small tail. Their relationship to downstream success is task-dependent. On AntMaze, where future goals align with path-like progress, these diagnostics explain landscape regimes. On Cube and Scene, goal ranking and manipulation control decouple: methods can rank goals well while failing downstream, or succeed through action-conditioned advantages despite weak future-vs-random separation. These results show that peak tuned success alone does not establish broadly extractable goal-conditioned behavior. Trainability landscapes expose this gap, while extraction diagnostics offer a lower-cost lens on how learned signals become policies.

Activation steering has emerged as a promising alternative for controlling language-model behavior at inference time by modifying intermediate representations while keeping model parameters frozen. However, large-scale evaluations such as AxBench show that existing steering methods are often outperformed by simple in-context prompting and generalize poorly to unseen concepts. We hypothesize that these limitations arise from unvalidated simplifying assumptions shared across prior methods, which typically restrict steering interventions to fixed, single-step, position-invariant transforms. We propose FLAS (Flow-based Activation Steering), which learns a general, concept-conditioned velocity field $v_t(h,t,c)$ that transports unsteered activations to steered ones without relying on these assumptions. On AxBench, FLAS is the first learned method to consistently outperform prompting, reaching held-out harmonic means of $1.015$ on Gemma-2-2B-IT and $1.113$ on Gemma-2-9B-IT without per-concept tuning. Analysis of the learned flow shows curved, multi-step, token-varying trajectories, which suggests that previous hypotheses on activation space geometry might be incomplete. Our code is available at https://anonymous.4open.science/r/FLAS.

Semi-supervised medical image segmentation has gained increasing attention for its outstanding performance with limited annotations. Most methods follow the Mean Teacher framework, where Exponential Moving Average (EMA) is primarily used to stabilize predictions, while its potential roles beyond stabilization are largely overlooked. In this work, we interpret EMA as a first-order infinite impulse response (IIR) low-pass filter with inherent frequency selectivity, which enables the preservation of semantic information at different levels. Motivated by this insight, we propose a dual-EMA teachers framework, including a long-term teacher focusing on global structures and a short-term teacher emphasizing local details. Furthermore, to help the student better exploit the guidance from both teachers, a First-In-First-Out replay queue is designed to reuse historical high-confidence pseudo-labeled pairs, leveraging pseudo labels provided at different training stages and reducing the sensitivity to sampling bias. Meanwhile, we impose feature-level global semantic and local relation consistency constraints to alleviate the impact of noisy pseudo labels, facilitating robust transfer of complementary representations from dual teachers to the student model. Experiments on three public medical image datasets demonstrate that our method outperforms current SoTA methods, verifying its effectiveness.

Dense correspondence matching has historically been bounded by simplifying spatio-temporal priors, such as smooth motion and rigid geometry. While effective for classical tasks, these foundations collapse at the frontier of Vision-Language-guided Image Editing and Generation (VL-IEG), where transformations yield correspondences that are perceptually obvious yet physically discontinuous. To transcend spatio-temporal priors, we draw inspiration from human cognition: learning a generalized model of identity-preserving visual consistency rather than relying on physical constraints. To achieve this, we introduce FreeMatching, a unified framework that extracts identity-preserving knowledge from foundation models and diverse datasets. The model's capability is forged through a two-stage training paradigm: supervised pre-training on diverse annotated datasets, followed by weakly-supervised refinement on data without dense annotations. Experimentally, FreeMatching not only achieves performance competitive with SOTA methods on classical benchmarks but also establishes the first strong baseline for the challenging VL-IEG task. Furthermore, we demonstrate its utility as a quantitative metric for evaluating identity-preserving consistency in generative editing, aligning closely with human judgment.

Contrastive vision-language (VL) pretraining in computational pathology has generally relied on curated image-caption datasets sourced from social media, educational videos, and research publications. These corpora are sparse relative to the diversity of tissue morphology and difficult to scale further, as pathologists do not routinely produce descriptive captions in clinical practice. We propose to instead leverage slide-level pathology reports which can be generated during routine diagnostic workflows in a scalable manner. A key challenge in this weakly-supervised setting is that pathology reports describe macroscopic, diagnostic findings while patch-level visual features describe microscopic findings. We introduce Locked Image Multiple Instance PreTraining (LIMIT), which addresses this gap by freezing a pretrained vision encoder and fine-tuning only the text encoder: report embeddings serve as cross-attention queries over bags of patch features, producing report-conditioned slide representations optimized via cross-entropy. Using LIMIT, we establish CONCH-Z, which aligns the text encoder of CONCH v1.5 to 335,645 clinical reports. Evaluated across 20 tasks spanning patch-level classification, tumor detection, and slide-level subtyping, CONCH-Z establishes state-of-the-art zero-shot performance over VL encoders trained on curated caption datasets, while simultaneously closing the gap with slide-level foundation models trained with substantially more parameters and compute. We will release CONCH-Z weights and evaluation code to support reproducibility and broader community use.

Conditional diffusion models have emerged as a powerful paradigm for image restoration, leveraging pre-trained generative priors to recover high-fidelity details from degraded inputs. However, existing methods typically follow a static inference protocol, treating the degraded observation as a fixed control signal throughout sampling. This overlooks the potential of inference-time scaling, where guidance can be refined on the fly to improve output quality. Recent test-time optimization approaches that manipulate noise variables or denoising trajectories often compromise structural fidelity, while those targeting text embeddings lack the spatial precision needed for restoration. To address this, we propose Condition-Aware Test-Time Optimization (CATTO), a training-free method that iteratively refines the visual conditioning signal at inference time. CATTO performs reward-aligned condition refinement under the pre-trained prior to enhance perceptual quality. To avoid complex backpropagation, we design a gradient-free optimization strategy guided by a joint objective: maximizing a perceptual reward while enforcing trajectory- and condition-level consistency to explicitly preserve structural fidelity. This process is accelerated by optimizing within a low-dimensional frequency subspace and reusing reliable updates across nearby denoising steps. Experiments on image restoration benchmarks show that CATTO improves perceptual quality over strong diffusion-based baselines without updating model parameters.

Multimodal Large Language Models (MLLMs) have demonstrated substantial promise in spatial understanding. Existing works typically incorporate prior knowledge extracted from a pre-trained foundation model to further enhance the spatial awareness of MLLMs. In this paper, we first reveal that when integrating diverse foundation models into MLLMs, each model provides complementary spatial priors that benefit different tasks. Motivated by this, we propose $\textbf{ViPS}$, a novel multi-model prior framework designed to fully unleash the potential of incorporating multiple $\textbf{Vi}$sual $\textbf{P}$riors from diverse models into MLLMs for $\textbf{S}$patial understanding. Specifically, ViPS introduces an Efficient Prior Proxy to generate multiple foundational priors with minimal inference overhead, and a Dynamic Prior Fusion mechanism to achieve harmonious and context-aware prior fusion and injection from the prior proxies. Extensive experiments demonstrate that ViPS successfully harmonizes diverse visual priors, establishing new state-of-the-art performance across multiple complex spatial reasoning and 3D spatial understanding benchmarks.

Modern retrieval-augmented systems typically access a corpus through a fixed similarity interface (either sparse or dense) that returns only the top-$k$ most similar documents for a query. While efficient, this design becomes a bottleneck for agentic search: exact lexical matching, conjunctions of weak clues, local context verification, and multi-step hypothesis refinement are difficult to realize with a conventional off-the-shelf retriever, and evidence filtered out early cannot be recovered downstream. This problem becomes particularly acute when agents must iteratively discover intermediate entities, combine partial evidence, and revise their search strategy. To address this problem, we study direct corpus interaction (DCI), a simple alternative in which an agent searches the raw corpus directly using general-purpose terminal tools (e.g., grep, file reads, shell commands, and lightweight scripts), without relying on embedding models, vector indexes, or retrieval APIs. This approach requires no offline indexing and adapts naturally to evolving local corpora. Across benchmarks for information retrieval (i.e., BRIGHT and BEIR) and end-to-end agentic search (i.e., BrowseComp-Plus), this simple setup substantially outperforms strong sparse, dense, and reranking baselines. It also attains strong accuracy on multi-hop question answering without relying on any conventional semantic retriever. Our results suggest that as language agents grow more capable, retrieval quality depends not only on reasoning ability, but also on how richly the corpus interface supports interaction. DCI therefore points to a broader interface design space for agentic search. Our code is available at https://anonymous.4open.science/r/dci-agent/.

Many decision-level fusion methods implicitly assume that semantically aligned soft outputs can be directly aggregated once they share the same label simplex. However, this assumption can fail even under identical label semantics, as oracle-specific differences in confidence sharpness, boundary uncertainty, and tail-mass allocation may place soft outputs in incompatible probability geometries and bias direct soft fusion. To address this issue, we formalize this failure mode as geometric incomparability and propose a pre-fusion comparability recovery framework, which learns a shared consensus through constrained oracle-specific rectifiers on the probability simplex. The framework includes a closed-form variant with frozen geometry descriptors and an alternating-refinement variant with consensus-dependent descriptors. Theoretically, we analyze the rectifier family, prove exact recovery under matched radial distortions, characterize the bias of direct fusion, and establish stable recovery under approximate mismatch. Experiments on controlled synthetic stress tests and real black-box decision-level settings show that comparability recovery consistently reduces fusion bias and improves reliability over direct aggregation. These results indicate that geometric comparability is a necessary condition for reliable soft-output fusion. The code is available at \url{https://anonymous.4open.science/r/CFCR-ACR-B8F1}.