NeurIPS 2026 论文列表

State-of-the-art vision-language models (VLMs) score impressively on video benchmarks yet stumble on basic visual reasoning tasks involving spatial relations, navigation, and object selection that a preschooler solves without effort. We hypothesize that the explicit pedagogical structure, specifically the context-question-pause-answer cycles embedded in children's educational video, provides naturally co-aligned reasoning traces: temporally synchronized visual cues, questions, and answers that emerge only from deliberate pedagogical authoring and cannot be practically reconstructed through manual annotation at scale. To test this, we introduce SoSVQA (Structure over Scale Visual Question Answering), a unified benchmark of 10K question-answer pairs automatically extracted from Dora the Explorer (DoraVQA) and Mickey Mouse Clubhouse (ClubHVQA) with precise timestamp alignment, and fine-tune Qwen2-VL and Qwen3-VL using Group Relative Policy Optimization (GRPO) to leverage the clear correctness signals and structured reasoning traces inherent in educational content. Despite training on just 10K QA pairs from 78 hours of children's television, orders of magnitude less data than GPT and Gemini, our approach delivers generalizable performance gains for Qwen-based VLMs, yielding consistent improvements on NExT-QA (+19.7), Video-MME (+10.6), and MotionBench (+4.9), matching the performance of leading proprietary systems and demonstrating that content structure can compensate for content scale.

Continuous DAG learners typically optimize a single weighted adjacency matrix, forcing each entry to represent both direct-edge support and signed effect magnitude. As a result, data fitting, sparsity, acyclicity, and any auxiliary structural information are all imposed on the same variables, even when the available information concerns higher-order relations such as reachability or ancestral order. We propose a structure-separated continuous DAG-learning framework that gives graph support its own soft representation. The effective adjacency is \( A=(W_f(M)\circ G(Z))\circ(\mathbf{1}-I), \qquad W_f(M)=\beta\tanh(M), \) where \(W_f(M)\) is a bounded signed-magnitude branch and \(G(Z)\) is a structural gate. Data fitting, sparsity, and the DAGMA barrier are evaluated on \(A\), while cycle and relation constraints are applied to closure scores computed from \(G\). The bounded magnitude branch prevents strong effective edges from being hidden behind vanishing structural gates, making closure-side structural regularization meaningful. On this separated structural gate, we introduce SMNC, a scan-efficient multi-scale normalized closure module. SMNC computes dyadic closure scores over \(G\) and normalizes each scale by its path-length budget, producing scale-normalized matrix-valued logits for self-reachability and off-diagonal reachability. This design addresses a key limitation of unnormalized product-style closure scores, whose path evidence scales as \(p^L\) on a length-\(L\) path with gate value \(p\), causing long-cycle signals to decay exponentially. At the appropriate covering scale, SMNC routes gradients through an active length-\(L\) cycle with a polynomial \(1/L\)-type scale factor; in the all-scale soft readout used in practice, this should be interpreted as a local active-scale effect. The same relation space supports diagonal self-reachability losses for cycle suppression and masked off-diagonal losses for optional ancestry, backward-exclusion, and incomparability cues, without treating positive reachability labels as direct-edge labels. Across linear and nonlinear SEM benchmarks, with ablations over factorization and closure normalization, we show that structure-separated SMNC improves DAG recovery and provides a principled closure-space interface for higher-order structural supervision.

Modern scientific discovery systems can propose more mechanistic hypotheses than laboratories can test. We study the downstream selection problem: after upstream retrieval, extraction, expert modeling, or curation has fixed a finite locally factorized posterior over admissible mechanisms, when does the posterior evidence justify acting on one structural hypothesis rather than its rivals? For a query $A\rightsquigarrow B$, we define *structural support* as the posterior mass of mechanisms in which $A$ reaches $B$ by a directed path. This is not a retrieval rank, direct-edge score, or MAP graph decision: support can be distributed across many compatible mechanisms. Our main result is a local certificate for this global support quantity. For path cutoff $L$ and locality radius $m$, truncated support is a convex combination of exact boundary-conditioned local supports. Optimizing over locally admissible blanket states gives a boundary range $[\\ell_{L,m},u_{L,m}]$, and a certified tail bound $\bar r_L$ accounts for longer witnesses: $p_{AB}\in[\\ell_{L,m},\min\{1,u_{L,m}+\bar r_L\}]$. For coupled rivals, the same boundary-conditioned calculations yield a joint feasible set rather than only independent coordinate intervals. We convert these certificates into resolve-or-abstain decisions: an action is resolved only when it wins throughout the certified support set. Under valid certificates, a resolved action is the unique strict utility maximizer inside the declared fixed model, action set, and utilities; otherwise, the solver abstains and reports whether boundary, tail, computation, evidence, or outer-model uncertainty blocks resolution. On controlled and biologically scaffolded finite-posterior records, the implementation has zero endpoint containment misses and zero false fixed-model resolutions. In a coupled-rival diagnostic, joint feasible sets resolve $14/24$ cases versus $6/24$ for scalar boxes, with no false commitments.

We propose Structural Self-Teaching (SST), a novel framework that enhances compositional image generation in unified multimodal models (UMMs) by leveraging their internal understanding path as a source of structural supervision. The motivation of this work is the observation that the generation path of UMMs often struggles with dense compositional prompts, leading to missing objects, attribute binding failures, and spatial errors. To bridge this gap, our key idea is to convert the phrase-level grounding inherent in the understanding path into training-time structural signals. Specifically, we extract phrase-conditioned support masks from internal attention maps derived from the understanding path and employ them through two complementary mechanisms: 1) phrase-local contrastive alignment to synchronize generated features with their corresponding phrase-level grounding, and 2) a lightweight structural scaffold for spatial modulation of intermediate generation states. By employing attenuated scaffold dropout during training and removing the scaffold branch at inference, our approach improves compositional fidelity across object, attribute, and spatial constraints. Crucially, the framework requires no external structural annotations for training or additional inputs at test-time, preserving the original generation interface. Extensive experiments across multiple UMMs demonstrate that our method consistently improves compositional fidelity while maintaining the efficiency of a lightweight tuning setup.

When distilling reasoning from large language models (LLMs) into smaller ones, teacher rationales for similar problems often vary wildly in structure and strategy. Like a chef who makes the same dish differently each time, this inconsistency burdens the student with noisy supervision that is hard to internalize. We propose Distillation through Reasoning Path Compression (D-RPC), which constrains the teacher to follow a compact, dynamically maintained bank of reusable high-level reasoning paths. For each training question, D-RPC retrieves the most relevant path and conditions the teacher to follow it, producing rationales that are consistent across similar problems yet diverse enough to cover different problem types. A PAC-Bayes analysis formalizes the resulting trade-off between bank size and coverage: smaller banks reduce supervision entropy but risk coverage gaps, and the generalization bound identifies an optimal intermediate size confirmed by our ablations. Across five math and commonsense reasoning benchmarks with two student models, D-RPC consistently outperforms chain-of-thought distillation, freeform rationale generation, direct distillation, and structured-supervision baselines, while using fewer tokens than template-heavy alternatives.

Communication learning is crucial for complex coordination in multi-agent reinforcement learning (MARL), yet existing methods often struggle with a fundamental dilemma between communication overhead and task performance. Typically designed without theoretical guidance, they produce unstructured and redundant communication graphs, limiting both efficiency and scalability. To address this, we propose Structural Entropy Optimized Communication (SEOC), a framework that directly optimizes inter-agent communication topology using the information-theoretic principle of Structural Entropy. SEOC guides the network to self-organize into a low-entropy state with sparse connectivity and distinct communities, enabling efficient and interpretable decentralized coordination. This is achieved through a unified mechanism that jointly enforces structural order and semantic conciseness, reducing topological complexity via a differentiable structural entropy loss while compressing messages with a node information bottleneck. Experiments on benchmarks such as the StarCraft Multi-Agent Challenge (SMACv2) demonstrate that SEOC significantly outperforms state-of-the-art methods.

We introduce structural causal bottleneck models (SCBMs), a novel class of structural causal models in which causal effects between high-dimensional variables are mediated by low-dimensional summary statistics, or *bottlenecks*. SCBMs provide a flexible framework for mechanism-specific, target-dependent dimension reduction while remaining estimable via standard learning algorithms. We prove that bottleneck variables are identifiable up to bijection from observational data, and validate this experimentally. We then study causal effect estimation in linear SCBMs under three standard adjustment strategies. For instrumental variables, we identify a population-level non-identifiability of naïve two-stage least squares when all variables are high-dimensional, and show that bottleneck representations restore identification. For mediation and backdoor adjustment, we prove that conditioning on bottleneck variables yields valid effect estimates with explicit finite-sample efficiency gains over conditioning on full high-dimensional variables. We argue that SCBMs provide a principled alternative to existing causal dimension reduction frameworks such as causal representation learning and causal abstractions.

Language exhibits inherent structures, a property that explains both language acquisition and language change. Given this characteristic, we expect language models to manifest their own internal structures as well. While interpretability research has investigated how models compute representations mechanistically through attention patterns and Sparse AutoEncoders, the organization of the resulting representations is overlooked. To address this gap, we introduce StructLens, a framework to analyze representations through a holistic structural view. StructLens constructs maximum spanning trees based on the semantic representations in residual streams, inspired by tree representation in dependency parsing, and provides summaries of token relationships in representation space. We analyze how contiguous tokens are also nearby in representation space and find that middle layers show the strongest local-span organization. Moreover, analysis of pre-training checkpoints reveals that smaller local units become detectable earlier in pre-training, and larger units later. Our findings demonstrate that StructLens provides insights into how models organize token representations across layers and training.

We study strongly adaptive regret for Online Convex Optimization (OCO) with time-varying movement costs, a setting that evaluates a learner's performance over arbitrary sub-intervals. This framework captures complex sequential decision-making problems where both the environment and the cost of switching decisions fluctuate over time. We propose a novel algorithm which, for any interval $[a,b] \subseteq [T]$, achieves a data-dependent regret bound of $ \widetilde{\mathcal{O}}\left( \sqrt{(D^2+DP_{[a,b]})\sum_{t \in [a,b]}(\\|g_t\\|^2+\lambda_t\\|g_t\\|)} \right) $, where $D$ is the diameter of the decision domain, $P_{[a,b]}$ is the path length of the comparator sequence over $[a,b]$, $g_t$ is the gradient of the loss, and $\lambda_t$ is the movement-cost coefficient at round $t$. Key to our approach is a novel reduction to sleeping experts with movement costs, a previously unexplored variant that is of independent interest. As a result of this development, we further establish the first strongly-adaptive guarantees under delayed feedback. Our approach also naturally extends to OCO with time-varying memory, where we obtain strongly-adaptive guarantees which improve the best-known bounds.

Knowledge distillation generally assumes a strong-to-weak relationship where stronger teachers yield better students. In this work, we examine this assumption about distillation in large language model (LLM) pretraining. By varying architecture sizes and training token budgets, we create strong-to-weak, same-level, and weak-to-strong teacher-student relationships. We study distillation's effectiveness under these relationships with different mixes between language and distillation loss. Three findings emerge: (1) with proper loss mixing, weak-to-strong and same-level distillation improves over standard pretraining, where even small and undertrained teachers benefit large students; (2) making the teacher stronger can lead to saturated or even reversed gains; (3) distillation improves generalization (out-of-domain, downstream) more readily than in-domain fitting. Our results provide practical guidance for choosing the teacher model with the proper loss for more effective LLM pretraining distillation.

Multi-modal large language models (MLLMs) achieve strong modality understanding by pairing a large language model (LLM) with an encoder for a target modality such as vision, video, or audio. However, improving an MLLM's capability for a given modality typically requires additional training on large modality-specific datasets, incurring substantial data collection and compute costs. Model merging offers an alternative, but it is often infeasible for data-scarce, large per-sample size, or domain-specific modalities (e.g., audio and video), where same-modality model variants are rarely available. In this work, we characterize an intriguing asymmetric phenomenon: merging a well-aligned, data-rich source-modality MLLM into a data-scarce target-modality MLLM substantially improves the target on its own benchmarks. Our theoretical and empirical analyses show that this gain stems from enhanced alignment between modality-specific and textual tokens, induced by the stronger donor modality. Specifically, we derive a mutual-information lower bound that is monotonic in alignment-related quantities and strongly correlated with downstream MLLM performance. Building on this principle, we propose Directional Cross-Modal Alignment Transfer (DCAT), a novel framework that transfers textual alignment from a strong, well-aligned source (donor) modality to a weak target (recipient) modality, boosting target-modality performance without further fine-tuning. We further show that the alignment-enhancing objective admits a closed-form weight-space solution computed from only a small calibration set. DCAT outperforms existing model-merging methods, offering an efficient path toward cross-modal alignment transfer.

Stroke-based rendering is often optimized as a pixel-level reconstruction task. This over-reliance on global error reduction not only degrades strokes into fragmented curves due to a lack of structured planning, but also allows locally defective primitives to be concealed by overlapping layers. To address these issues, we introduce a model-guided framework for stroke-audited rendering. Our framework uses Quadratic Bezier Splatting to analytically model continuous curved strokes with efficient differentiable alpha compositing. Built on this renderer, a feedforward multi-scale stroke decoder provides a structural initialization with learned stroke-scale allocation, while test-time stroke optimization supplies the reconstruction capacity needed for high-fidelity refinement. We further propose a single-stroke audit mechanism that explicitly evaluates local stroke quality, preventing defective primitives from surviving optimization through layer-wise compensation. Experiments demonstrate our approach achieves highly coherent layouts, improving PSNR by up to 20\% and yielding a 58.8$\times$ renderer speedup with a 16\% more compact representation compared to representative SBR baselines.

Neural network verifiers aim to provide formal guarantees on model behavior, but existing verification benchmarks are fundamentally limited by their lack of ground-truth labels. As a result, verifier evaluation relies on indirect heuristics, which prevents exact scoring and systematic studies of verifier failure modes. We address this gap by introducing a reusable framework for generating verification instances whose ground-truth robustness labels are known a priori through analytic construction. Our framework led to the discovery of multiple numeric tolerance concerns and implementation bugs in popular verifiers, highlighting the need for ground-truth labels. Additionally, to systematically study verifier failure modes, we introduce the verification Difficulty Profile, a collection of estimable quantities capturing distinct sources of instance hardness. Using our framework and these profiles, we evaluate five state-of-the-art verifiers and show that different instances stress distinct aspects of the verification pipeline. We show that these results can aid the future development of verifiers as they provide actionable targets for improving numerical reliability, relaxation quality, and search behavior. Our code is publicly available: https://github.com/NeurIPSAnon12345/VeriStressGT.git.

As LLM agents advance rapidly in autonomy, the risk of hidden objectives that bypass safety constraints becomes increasingly plausible. Chain-of-thought (CoT) monitoring offers a key control mechanism by reading the agent's reasoning and flagging misalignment before it materializes into action, yet its reliability depends on the realism of the red-team it is stress-tested against. Existing red-teaming efforts apply uniform attack templates that treat the side task as an exogenous insertion, yielding abrupt anomalies any reasonable monitor can flag while leaving the subtler, contextually grounded evasions of a capable agent largely unprobed. We therefore ask: How robust is CoT monitoring against agents that strategically conceal misalignment? To answer this, we introduce TraceMRT (Trace-level Monitor Red-Teaming), an automated red-teaming framework that reframes the side task as a coherent sub-goal of the primary task and searches trace-shaping strategies whose induced trajectories complete the misaligned objective while reading as natural execution. Through black-box attacks on state-of-the-art monitors (e.g., GPT-5.2, and Claude-4.5), we reveal substantial vulnerabilities even in top models. To better counter these vulnerabilities, we further improve monitoring scaffolding with Bayesian Trajectory Monitoring, integrating global context and local evidence to better detect covert misalignment. Empirically, the method significantly consolidates the monitor with 0.91 AUC and 0.74 TPR@FPR=0.05.

Tabular prediction has long been dominated by gradient-boosted decision trees and specialized deep tabular models, while large language models (LLMs) remain difficult to make competitive despite their cross-task adaptability and transparent reasoning traces. Existing LLM-for-tabular methods mainly adapt tables into textual inputs, but often fail to capture key structural properties of tabular data. In this work, we propose Permutation Relative Policy Optimization (PRPO), a reinforcement learning post-training method that strengthens LLMs for tabular prediction by injecting column-permutation invariance as a structural prior. By constructing label-preserving column permutations and estimating advantages both within and across them, PRPO converts sparse outcome rewards into denser and more stable optimization signals. Extensive experiments on 139 OpenML datasets show that our 8B model reaches a genuinely competitive regime against strong specialized tabular baselines. It achieves strong fully supervised performance, dominates cross-dataset zero-shot settings, and performs on par with 32-shot strong baselines. Moreover, it substantially outperforms much larger general-purpose and reasoning LLMs, including up to a 53.17% improvement over DeepSeek-R1 (685B). These results show that structural-prior RL post-training is an effective route for making LLMs competitive in tabular prediction.

Deep neural networks have achieved significant success on in-distribution (ID) tasks, but out-of-distribution (OOD) samples during inference can lead to unreliable predictions. To address this issue, distance-based methods exploit the geometric structure of feature space for OOD detection. However, methods based on static information geometry fail to address geometry distorted by ill-distributed samples. Although prior work reduces the misclassification of OOD samples as ID by adjusting the residual space with real-time features, it fails to address the misclassification of ID samples as OOD. In addition, existing distance-based methods ignore class-specific characteristics when computing deviation feature and rely on the nearest prototype for uncertainty estimation, leading to suboptimal OOD detection performance. To address these issues, we propose Adaptive Mahalanobis Gap (AMG). Specifically, we first strengthen the principal space along real-time feature directions to reduce the misclassification of ID samples as OOD. Subsequently, we incorporate class variance and prototype norms to mitigate bias of deviation feature from class-wise distributional differences. Finally, we introduce the Mahalanobis gap scoring for uncertainty scoring to overcome bias from single-prototype reliance. Experiments on CIFAR-10/100 and ImageNet-1k show that AMG consistently outperforms state-of-the-art methods in FPR@95 and AUROC. Our code is available at https://anonymous.4open.science/r/AMG_AR.

Learning everyday skills, like cooking a dish, relies increasingly on instructional media such as online videos. This opens the door to the use of video (and multimodal) large language models (LLMs) as task guidance assistants. A crucial capability of the real-world success of a prospective task guidance assistant is its ability to intervene proactively as soon as a mistake is apparent in order to guide the user. To evaluate this crucial capability, we introduce Ego-MC-Bench (Mistake Corrections), a benchmark for evaluating *reactive, step-by-step* task guidance in realistic cooking scenarios. Extensive experiments show that Ego-MC-Bench is highly challenging for state‑of‑the‑art video LLMs. We argue that a key reason is the limited availability of training data for fine-tuning models on this task. Although there exists a wide range of cooking video datasets, existing datasets lack examples of mistakes along with appropriately timed interventions. To help address this data limitation, we also introduce Ego-CoMist, a counterfactual synthetic dataset created by transforming non‑interactive cooking videos into supervised training examples showing proactive interventions. We show that fine-tuning on Ego-CoMist yields performance gains especially for smaller and more efficient video LLMs that are well suited for delivering assistance on edge devices.

Inferring the evolution of high-dimensional and multi-modal (e.g., spatio-temporal) physical fields from irregular sparse measurements in real time is a fundamental challenge in science and engineering. Existing approaches, including diffusion-based generative models and functional tensor methods, typically operate in offline settings, depend on full temporal observations, or incur substantial inference cost. We propose StreamPhy, an end-to-end framework that enables efficient and accurate streaming inference of full-field physical dynamics from incoming irregular sparse measurements. The framework integrates a data-adaptive observation encoder that is robust to arbitrary observation patterns, a structured state-space model that supports memory-efficient online updates across irregular time intervals, and an expressive Functional Tensor Feature-wise Linear Modulation (FT-FiLM) decoder for continuous-field generation. We prove that FT-FiLM is more expressive than the functional Tucker model, admitting a richer function class for handling complex dynamics. Experiments on three representative physical systems under challenging sampling patterns show that StreamPhy consistently outperforms state-of-the-art baselines, with at least 48\% improvement in accuracy and up to 20--100x faster inference than diffusion-based methods.

Vision-Language-Action (VLA) models have demonstrated effectiveness in robot manipulation, yet state-of-the-art models such as $\pi_{0.5}$ operate under a single-frame paradigm, limiting their ability to retain past observations and develop precise spatial perception. In this paper, we propose StreamPI, a streaming multimodal temporal modeling framework that equips single-frame VLA with temporal reasoning capability without introducing any additional parameters. One core design is instruction-anchored temporal modeling. It treats each (visual observation, language instruction) pair as an atomic temporal unit: bidirectional attention within each pair enables cross-modal fusion, while causal attention across pairs preserves autoregressive streaming inference. This ensures the language instruction serves as a persistent semantic anchor throughout task execution. To bridge the gap between synchronous training and asynchronous real-robot deployment, we introduce a random-interval streaming training strategy: a proper inter-frame interval (e.g., every 3 frames) enables faster and smoother action execution. Beyond this, randomizing the interval further improves robustness to frame-timing perturbations, supporting asynchronous deployment in practice. Furthermore, by leveraging the length extrapolation capability of the LLM backbone, StreamPI seamlessly inherits pretrained single-frame weights and supports flexible single-frame and multi-frame inference. Experiments on real-robot tasks spanning memory-dependent and precise perception scenarios, as well as the simulation benchmark LIBERO, demonstrate that StreamPI outperforms $\pi_{0.5}$ across diverse tasks.