NeurIPS 2026 Papers

Robot manipulation often depends on short-term cues—such as brief contacts, temporary occlusions, and intermediate object states—that are ambiguous in a single frame and can only be resolved by integrating observations over time. However, injecting historical information solely by enlarging the history window is computationally costly and frequently yields limited gains due to substantial redundancy and task-agnostic, inflexible sampling of observations. We propose CondenseVLA, a multi-frame VLA policy with a lightweight learnable history condensation module for using short-term visual context in action decoding. Given a sequence of cached per-frame features from the VLM backbone, CondenseVLA distills observation history into a fixed set of slot tokens using learnable queries and multi-head cross-attention. Simple auxiliary regularization encourages diverse attention patterns by penalizing overlap among slot attention maps and reduces slot 12 redundancy by decorrelating slot embeddings, stabilizing learned condensation without imposing strong time priors. By expanding the history window and distilling it into a compact token set, CondenseVLA improves over the baseline on the evaluated Simpler-WidowX tasks, increasing success rate from 58.4 to 70.8 (+12.4), achieves 97.2% on LIBERO, and reaches 52.0% on real-world tasks. Beyond the gains, our results show that long history windows contain substantial redundancy, so naively providing more frames yields limited benefits, whereas learned condensation into fixed slot tokens provides an effective bounded-token way to leverage temporal context.

Generative ranking models (GRMs) have emerged as a scalable alternative to traditional recommendation pipelines by unifying items and user actions into a single token sequence and formulating recommendation as next-token prediction. Despite their promise, we identify a geometric failure mode: attention interactions with lowcardinality action tokens can induce dimensional collapse in high-rank item representations, as suggested by interaction-collapse theory. To pinpoint the origin of collapse, we view attention mechanism over item–action interleaving sequences as a mixture of directional interaction channels between token types, and find that collapse concentrates in the action-attending-item channel, where low-rank action gating drags the item value stream toward a low-dimensional subspace. To address this, we theoretically show that nonlinearity mitigates collapse, and propose Selective VAlue Nonlinearity (SVAN), a simple yet effective approach that applies nonlinear activation to the value component in the collapsed channel, protecting item representations while preserving the remaining attention geometry. SVAN consistently improves the ranking quality over strong baselines. Further analyses confirm that SVAN alleviates dimensional collapse, yielding up to a 12.58% effective-rank gain, and enhances representation discriminability.

Learning to Defer (L2D) enables a model to predict autonomously or defer to an expert, but prior work largely assumes flat label spaces. We study the first L2D setting with hierarchical multi-label decisions, motivated by medical-imaging workflows in which findings are organised by clinical taxonomies. In this setting, deferral is a delegation action rather than a label assignment, so treating it as an independent per-label decision can produce \emph{deferral incoherence}, including taxonomic contradictions, delegation violations, and deferrals of labels already implied by the model's own assertions. We formalise coherent hierarchical deferral under a Selective-Exclusion handoff contract, characterise the Bayes-optimal coherent deferral rule, and show that even nodewise Bayes L2D can be action-incoherent. We then propose two remedies: exact coherent projection, a dynamic-programming decoder over the coherent action set, and Taxonomic Belief Propagation (TBP) with Recursive Policy Optimisation (RPO), a contract-aware joint action model trained through the same recursion used at inference. Across real-reader and controlled-expert medical-imaging benchmarks, naïve binary-relevance L2D exhibits non-trivial incoherence. Projection removes it exactly, and fast TBP+RPO drives incoherence near zero while retaining strong utility.

As LLM agents advance beyond short-horizon tasks, evaluating their ability to perform long-horizon strategic decision-making in realistic multi-agent settings remains a key challenge. Existing agent benchmarks typically focus on isolated task completion or interactions with rule-based or fixed-policy counterparts, limiting their ability to capture strategic adaptation and emergent behaviors arising from sustained multi-agent interactions. We introduce \textbf{CoffeeBench}, a benchmark for evaluating LLM agents in a dynamic multi-agent economic environment consisting of farmers, roasters, and retailers forming a multi-stage supply chain. The environment consists of autonomous firms interacting over a multi-month horizon with evolving supply, demand, and pricing, creating sustained competition and supply-chain dynamics. In each evaluation, the evaluated model controls one firm and must make sequential decisions on procurement, production, pricing, and negotiation to maximize cumulative net income. CoffeeBench enables the study of emergent behaviors arising from repeated strategic interactions, including negotiation dynamics, pricing discipline, and coordination failures. We evaluate several frontier and pareto-frontier LLMs on CoffeeBench to analyze their economic performance and strategic behavior. Most evaluated models achieve positive net income, with GPT-5.5 achieving the highest mean net income. In contrast, Claude~Haiku~4.5 frequently yields negative net income, exhibiting an idle-drift failure mode in which agents collapse into inactivity despite coherent plans. We further find that stronger models proactively negotiate with counterparties through frequent messaging, whereas weaker models trade reactively with limited communication. CoffeeBench provides a controllable testbed for studying strategic interaction, coordination, and long-horizon behavior in multi-agent economies.

We study a canonical multi-task demand-learning problem motivated by retail pricing, where a firm seeks to estimate heterogeneous linear price-response functions across multiple decision contexts. Each context is described by rich covariates but exhibits limited price variation, motivating transfer learning across tasks. A central challenge in leveraging cross-task transfer is endogeneity: prices may be arbitrarily correlated with unobserved task-level demand determinants across tasks. We propose a new meta-learning framework that identifies the conditional mean of task-specific causal demand parameters given a subset of task-specific observables despite such confounding, assuming that each task contains at least two distinct locally exogenous price points. This subset is carefully designed to include all of the prices to address cross-task confounding, while masking two demand outcomes that provide randomized supervision to address identifiability issues arising from the inclusion of all prices. We show that this information design is maximally uniformly valid, in that any refinement of the conditioning set that reveals withheld-outcome information is not guaranteed to identify the conditional mean causal target. We validate our method on real and synthetic data, demonstrating improved recovery of demand responses relative to standard transfer-learning baselines.

Source labels tell us where a sample came from, but not how that source oriented its coordinates. In a sparse Gaussian multi-source model with source-specific sign gauges, that distinction changes the learning problem. Without calibration, pooling across sources erases the mean signal, and the source-labelled experiment reduces to sparse PCA. With task-independent relative-gauge metadata, synchronization aligns sources and the same observations become a sparse mean problem. We formalize the value of metadata through aligned mass, which quantifies how much first-order signal a metadata-induced alignment restores. Source labels create only limited aligned mass, whereas valid calibration can create aligned mass proportional to the full sample size. This yields a statistical separation in micro-source regimes and, under standard sparse-PCA hardness assumptions, a computational separation as well. Experiments confirm the predicted phase transition, the failure of invalid metadata, reuse of one calibration graph across many tasks, and controlled real-feature settings with imposed source gauges.

3D human pose estimation from sparse multi-view camera rigs is an essential task for numerous applications, including action recognition, sports analysis, and human-robot interaction. While learned methods dominate the field on benchmarks, they require large annotated datasets; training-free optimization-based methods remain promising as they circumvent 3D supervision by solving a correspondence problem across views from 2D detections. Existing combinatorial formulations rely on pairwise associations to model this correspondence problem and enforce global consistency across views only as a downstream constraint. However, reconciling locally plausible pairwise matches becomes brittle under occlusion and noisy detections, where local errors propagate globally. We propose COMPOSE, which recasts multi-view 3D human pose estimation as a weighted exact-cover optimization over a hypergraph of person hypotheses. Our formulation replaces pairwise association and post-hoc consistency enforcement with a single global combinatorial objective. To address the exponentially large candidate space, we introduce a geometric pruning strategy alongside two complementary solvers: an exact Integer Linear Programming formulation and a scalable relaxation via Belief Propagation. Without any 3D supervision, COMPOSE improves average precision by up to 31 points over the best optimization-based method and 13 points over self-supervised learned methods, demonstrating the effectiveness of higher-order combinatorial association for training-free multi-view 3D human pose estimation.

Off-policy evaluation (OPE) estimates the value of a target policy from logged data, but limited behavior-policy coverage can force high-variance reweighting or reward-model extrapolation. Counterfactual annotations can add evidence about unobserved actions, yet practical sources, including domain experts and LLMs, may be costly, biased, or noisy. We study budgeted acquisition of such annotations for contextual-bandit OPE. Given source-specific costs and error profiles, we formulate an integer allocation problem over context-action pairs and annotation sources to minimize the component of estimator variance that depends on the annotation plan. We characterize when annotations are valuable through a first-annotation threshold and local annotation-value regimes. For the coupled multi-source problem, we develop a majorization-minimization algorithm with dynamic-programming subroutines that monotonically improves the objective. Experiments in synthetic clinical and LLM-annotated education bandits show that our proposed allocation method reduces the optimized variance component by $20.58$% and $10.77$%, respectively, relative to no-annotation baseline.

While standard reinforcement learning optimizes a single reward signal, many applications require optimizing a nonlinear utility $f(J_1^\pi,\dots,J_M^\pi)$ over multiple objectives, where each $J_m^\pi$ denotes the expected discounted return of a distinct reward function. A common approach is concave scalarization, which captures important trade-offs such as fairness and risk sensitivity. However, nonlinear scalarization introduces a fundamental challenge for policy gradient methods: the gradient depends on $\partial f(J^\pi)$, while in practice only empirical return estimates $\hat J$ are available. Because $f$ is nonlinear, the plug-in estimator is biased ($\mathbb{E}[\partial f(\hat J)] \neq \partial f(\mathbb{E}[\hat J])$), leading to persistent gradient bias that degrades sample complexity. In this work we identify and overcome this bias barrier in concave-scalarized multi-objective reinforcement learning. We show that existing policy-gradient methods suffer an intrinsic $\widetilde{\mathcal{O}}(\epsilon^{-4})$ sample complexity due to this bias. To address this issue, we develop a Natural Policy Gradient (NPG) algorithm equipped with a multi-level Monte Carlo (MLMC) estimator that controls the bias of the scalarization gradient while maintaining low sampling cost. We prove that this approach achieves the optimal $\widetilde{\mathcal{O}}(\epsilon^{-2})$ sample complexity for computing an $\epsilon$-optimal policy. Furthermore, we show that when the scalarization function is second-order smooth, the first-order bias cancels automatically, allowing vanilla NPG to achieve the same $\widetilde{\mathcal{O}}(\epsilon^{-2})$ rate without MLMC. Our results provide the first optimal sample complexity guarantees for concave multi-objective reinforcement learning under policy-gradient methods.

Large language model (LLM) agents increasingly automate machine-learning model building, yet biomedical benchmarks emphasize question answering, reasoning, and tool use or cover narrow slices of biomedical ML coding. We introduce BioXArena, a biomedical machine learning (BioML) coding benchmark for task-specific model building on heterogeneous, often multi-modal datasets. It contains 76 end-to-end tasks across 9 domains: sequence, single-cell, structure, network biology, chemical biology, perturbation dynamics, phenotype--disease, imaging, and text-integrated tasks. Each task packages primary-source data with hidden labels, held-out graders, and biology-aware metrics on a unified 0-to-100 scale; agents write runnable code, train models, and predict private test samples. BioXArena emphasizes realistic data interfaces: most tasks require agents to integrate multiple input sources, including 46 tasks that combine distinct modality families. We compare 18 harness--backbone configurations: general-purpose coding, biomedical, and ML agents plus our Minimal runner, all under the same controlled resource configuration. Codex with GPT-6 Astra achieves the highest task-level average score of 73.91 among 18 configurations, closely followed by Claude Code with Opus 5.5 (73.79). Matched-backbone comparisons show higher observed scores for mature coding systems than for Minimal; seven Minimal backends characterize model differences under a fixed runner. Beyond the main leaderboard, ablations, robustness checks, and scaling, runtime, and failure-mode analyses characterize how backbones, scaffolds, budgets, and domains affect BioML coding performance, exposing modeling and execution gaps across agent and LLM backbone families.

While LLMs excel at single-turn generation, they struggle with long-horizon, multi-turn interactions. Offline reinforcement learning (RL) offers a scalable approach, yet its performance hinges on the availability and quality of multi-turn trajectory data. A common remedy is to augment training with synthetic trajectories generated by LLMs or simulators, but synthetic data is highly heterogeneous in quality, and naively treating all trajectories as equally informative can degrade performance. We propose BOOST, a bilevel optimization framework where the inner level trains the LLM on reweighted data and the outer level trains a lightweight reweighting head on held-out real validation tasks, assigning continuous trajectory-level weights without requiring an external judge. To ground this approach, we derive a PAC-Bayesian bound revealing a three-way trade-off: synthetic data increases diversity but risks task-shift, while concentrating weight on high-quality trajectories improves empirical performance at the cost of effective sample size. Empirically, our method consistently outperforms multiple baselines. Analysis reveals it upweights synthetic trajectories that align with the real data distribution and exhibit higher qualitative merit.

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.

Automatic Prompt Optimization (APO) aims to improve the capabilities of Large Language Models (LLMs), yet existing methods remain limited by their result-oriented nature. Training-free approaches often converge to suboptimal prompts, while training-based methods suffer from reward sparsity and optimization instability. More importantly, both paradigms focus on obtaining the final prompt rather than learning the prompt optimization process itself. To address this dilemma, we propose \textbf{ProPO}, a trajectory-driven \textbf{Pro}cess learning framework for \textbf{P}rompt \textbf{O}ptimization guided by multi-dimensional composite rewards. Specifically, ProPO introduces a historical trajectory reflection mechanism for high-quality exploration. Concurrently, we adopt adaptive rubric rewards to provide fine-grained supervision and alleviate reward sparsity. More importantly, we design a novel Process Logic Reward that supervises the consistency of the reflection process. Extensive experiments demonstrate that ProPO, integrated with GRPO, establishes a new SOTA performance. Notably, ProPO achieves a remarkable \textbf{70.00\%} accuracy on the highly complex AIME-2026 task and an absolute gain of \textbf{+25.25\%} over the zero-shot baseline on Subjectivity Classification (Subj). Furthermore, empirical analyses reveal that our multi-dimensional reward mechanism provides finer-grained and denser signals, allowing the model to learn the prompt optimization process, and ultimately enabling it to generate significantly higher-quality prompts.

Offline evaluation of ad ranking policies in multi stage delivery systems is fundamentally challenging because action propensities are often unavailable, limiting inverse propensity scoring (IPS) estimators. While reward model based estimators such as Direct Method (DM) do not rely on propensity estimates, they can suffer from bias when candidate policies induce impression distributions that differ from the logging mixture. In this work, we formalize DM based counterfactual replay under mixture logging and propose a mixture targeted shift correction based on density ratio weighting through domain classifier. We further derive a per policy value estimation error and characterize an asymptotic error ceiling governed by the cold start mass. Experiments on both ads delivery simulator and one week of production traffic from a large scale industrial CTR ranking system demonstrate that the proposed mixture weighted reward model consistently outperforms its unweighted counterpart under distribution shift.

Early-exit networks promise inference savings by terminating computation at intermediate classifiers, but FLOPs-based gains rarely translate into on-device latency reduction under static-graph compilation. On NVIDIA Jetson Orin Nano, the per-sample exit policy of Shallow-Deep Networks (SDN) runs approximately 2× slower than vanilla ResNet-56 and misses the 60 fps deadline despite using only 40% of the FLOPs, because dynamic branching precludes layer fusion. We propose the Selective Lightweight IC Ensemble Network (SLIENet), a train-full, deploy-partial framework compiled as a single static FP16 engine. SLIENet trains the full backbone with light self-distillation, using the final classifier as a stop-gradient teacher to preserve error diversity across internal classifiers (ICs). At deployment, a sub-second calibration search selects an IC subset for softmax averaging, while overthinking late blocks are truncated at depth k. On CIFAR-100 across ResNet-56, VGG-16, and MobileNetV1, SLIENet outperforms SDN-based variants and remains competitive with ZTW. On Jetson Orin Nano, the recommended SLIE5 configuration improves accuracy by +2.46 pp while reducing p50 latency by 14% and energy by 15% over vanilla, with zero deadline misses across 10,000 test samples. Thus, multi-exit deployment should be a pre-compiled deterministic inference plan, not a sample-wise routing policy.

As foundation models scale toward fusing more heterogeneous visual streams, understanding how diverse encoders interact under joint training becomes a prerequisite for principled design. Yet large vision-language models (LVLMs) currently lack the tools to do so, and parameter-efficient encoder configurations remain hard to identify before training. To re-examine encoder roles under joint training, on the 16-benchmark Cambrian-1 suite we retrain and evaluate all 31 non-empty subsets of five common vision encoders under a unified pipeline ($\approx$20k GPU-hours total), and report three findings. First, retraining each subset from scratch reveals encoder rankings that differ from those obtained by masking encoders on a fixed checkpoint, including which encoder ranks first overall. Second, we decompose each encoder's contribution into two axes, \emph{Capacity}, the score an encoder reaches on its own, and \emph{Necessity}, the drop when it is removed from the full pool. The two axes are not interchangeable. Pairing the two highest-Capacity encoders is suboptimal, while pairing a high-Capacity anchor with an adaptive complement matches the full five-encoder model. Adding further encoders beyond this pair yields only marginal gains. Third, at fixed parameter count, per-encoder pre-projector effective rank explains the residual score variation. The strongest pairs combine an anchor whose rank survives joint training with a complement whose rank \emph{expands} under it, suggesting that higher-rank, less-collapsed projector inputs correspond to a more favorable optimization regime at the encoder–projector interface. Together, the Capacity–Necessity decomposition and the pre-projector rank analysis, along with comprehensive evaluation through retraining, expose a methodological gap in multi-encoder LVLM design, and offer concrete primitives for closing it.

Multi-task learning requires a shared update that makes balanced progress across multiple objectives. A key obstacle is that progress is difficult to compare across tasks with different loss scales, units, and local optimization dynamics. We study this problem from an optimality-gap perspective. For each task, we compare the improvement induced by the shared update with the best one-step improvement achievable by optimizing that task in isolation. This comparison yields a scale-invariant normalized improvement rate, which measures how well the shared update serves each task relative to its own local improvement bound. We formulate balanced multi-task optimization as a max-min problem over the normalized improvement rates, thereby prioritizing the worst-served task. A local quadratic approximation leads to the Karush-Kuhn-Tucker (KKT) optimality conditions, showing that the optimal shared update is determined by a small set of bottleneck tasks. Under an isotropic Hessian approximation, the update admits a closed-form expression that depends only on task gradients and a regularized Gram matrix, while the bottleneck set is identified by an active-set procedure. Experiments on diverse benchmarks show improved task balance and competitive performance.

Constrained preference-conditioned MORL typically assumes full task observability, treating preferences and safety budgets as stationary task signals. In real-world deployment, however, these signals are often unreliable proxies: intent may drift, safety tolerances may shift with context, and exposed descriptors may be noisy, delayed, or incomplete. We therefore formulate constrained preference-conditioned MORL as a partially observed task-specification problem, where the agent must navigate a fundamental conflict between unreliable external signals and latent interaction dynamics. We propose \textbf{BECON}, a belief-conditioned framework for constrained MORL with drifting preferences and budgets, which infers latent task context from recent histories to estimate preferences and safety budgets. These estimates are fused with the exposed task signal via uncertainty-dependent gates, enabling the policy to defer to observations under high epistemic uncertainty while correcting them when history is informative. The fused context jointly conditions preference optimization and context-adaptive constraint enforcement. Experiments on diverse multi-objective control tasks demonstrate that BECON consistently improves over state-of-the-art MORL baselines, achieving stronger robustness and constraint satisfaction under task drift.

Understanding how distributed brain regions coordinate during behavior requires models that are both predictive and interpretable. We introduce Behavior-Adaptive Connectivity Estimation (BACE), an end-to-end framework for learning behavior-conditioned directed effective connectivity from multi-region intracranial local field potentials (LFPs). BACE first encodes within-region dynamics with region-specific temporal encoders, then applies a learned adjacency matrix selected for each behavioral context, and finally forecasts future neural activity through a graph-conditioned autoregressive decoder. This design yields explicit region-level connectivity matrices whose edges are tied to predictive dynamics rather than post-hoc explanation. On controlled synthetic time series with known directed graphs, BACE recovers the ground-truth edge structure from forecasting alone. We then evaluate BACE on human deep-brain LFP recordings from three participants performing structured motor tasks. Across participants, BACE achieves strong neural forecasting while producing compact, behavior-specific directed graphs that can be inspected across task segments. Reliability analyses further support the stability of the inferred connectivity patterns. Together, these results position BACE as a practical framework for estimating behavior-adaptive effective connectivity from high-dimensional intracranial recordings, enabling interpretable hypotheses about how deep-brain networks reorganize during behavior.