ICML 2026 论文列表

Process reward models (PRMs) provide fine-grained supervision for reasoning, but reliable PRMs often require step annotations or heavy verification pipelines, making them costly to scale and refresh during online RL. Implicit PRMs reduce this cost by training log-likelihood-ratio rewards from trajectory-level outcome labels. However, the log-ratio is constrained only as a sequence-level aggregate during training, while inference decomposes it into token- or step-level scores for partial prefixes. This train–inference mismatch leaves local credits weakly identified, so distribution-wide scoring can amplify misleading advantages. We propose Implicit Prefix-Value Reward Model (IPVRM), which directly learns the probability of eventual correctness for each prefix from outcome labels. Step signals are then obtained as temporal-difference (TD) differences between consecutive prefix values, aligning the training target with inference-time use. IPVRM markedly improves step-verification F1 on ProcessBench. To exploit these prefix values during policy optimization, we further introduce Distribution-Level RL (DistRL), which applies TD advantages to both sampled tokens and high-probability candidate tokens, providing dense counterfactual updates without additional rollouts. Experiments show that DistRL brings limited gains with unreliable implicit rewards, but consistently improves downstream reasoning when paired with IPVRM. The implementation of our method is available at https://github.com/gaoshiping/IPVRM .

It is increasingly important that LLM agents interact effectively and safely with other goal-pursuing agents, yet, recent works report the opposite trend: LLMs with stronger reasoning capabilities behave _less_ cooperatively in mixed-motive games such as the prisoner's dilemma and public goods settings. Indeed, our experiments show that recent models---with or without reasoning enabled---consistently defect in single-shot social dilemmas. To tackle this safety concern, we present the first comparative study of game-theoretic mechanisms designed to enable cooperative outcomes between rational agents _in equilibrium_. Across four social dilemmas testing distinct components of robust cooperation, we evaluate four families of mechanisms: (1) repeating the game for many rounds, (2) reputation systems, (3) third-party mediators to delegate decision making to, and (4) contract agreements for outcome-conditional payments between players. Among our findings, we establish that contracting and mediation are most effective in achieving cooperative outcomes between capable LLM models, and that repetition-induced cooperation deteriorates drastically when co-players vary. Moreover, we demonstrate that the mechanisms become _more effective_ under evolutionary pressures to maximize individual payoffs.

ANN-to-SNN conversion offers energy-efficient inference but faces a fidelity-latency trade-off due to open-loop error accumulation. While conversion-aware training mitigates this, it sacrifices the generality of using off-the-shelf ANNs. We propose a closed-loop fine-tuning framework that calibrates these errors without altering the source model. Our approach employs a Dual Alignment Mechanism, utilizing global Kullback-Leibler divergence for output distillation and introducing an adaptive local Centered Kernel Alignment constraint, weighted by initial conversion loss, for feature alignment. We uncover a critical time-dependent dynamic: local constraints are essential for stabilizing representations in low-latency regimes (e.g., $T=8$) where global gradients are unstable, whereas global alignment drives fidelity at higher time steps. Experiments on CIFAR-10 demonstrate that our method achieves over 99\% of source ANN accuracy at $T=32$ (e.g., ResNet-18: 96.38\% vs.\ 96.39\%). Furthermore, this fine-tuning acts as a regularizer, yielding SNNs with input noise robustness that matches or exceeds the source ANN.

Federated Learning (FL) enables collaborative training of distributed clients while protecting privacy. To enhance generalization capability in FL, prototype-based FL is in the spotlight, since shared global prototypes offer semantic anchors for aligning client-specific local prototypes. However, existing methods update global prototypes at the prototype-level via averaging local prototypes or refining global anchors, which often leads to semantic drift across clients and subsequently yields a misaligned global signal. To alleviate this issue, we introduce **hyper-prototypes**, defined by a set of learnable global class-wise prototypes to preserve underlying semantic knowledge across clients. The hyper-prototypes are optimized via gradient matching to align with class-relevant characteristics distilled directly from clients' real samples, rather than prototype-level descriptors. We further propose **FedHPro**, a Federated Hyper-Prototype Learning framework, to leverage hyper-prototypes to promote inter-class separability via mutual-contrastive learning with client-specific margin, while encouraging intra-class uniformity through a consistency penalty. Comprehensive experiments under diverse heterogeneous scenarios confirm that 1) hyper-prototypes produce a more semantically consistent global signal, and 2) FedHPro achieves state-of-the-art performance on several benchmark datasets.

Video action models are a promising foundation for Vision–Language–Action (VLA) because they can learn rich visual dynamics directly from video. However, likelihood-oriented training of diffusion predictors emphasizes globally plausible futures and does not guarantee precision-critical visual dynamics needed for manipulation, so small prediction errors can be amplified by downstream policies. We propose Dyn-VPP, a post-training framework that casts multi-step denoising as policy optimization and aligns predicted future latents with expert visual dynamics via verifiable terminal reward, without modifying any architecture. This enables explicit optimization of dynamics signals that are not captured by likelihood-only training. As a result, Dyn-VPP yields more accurate visual dynamics and improves downstream task execution. Experiments across diverse simulated and real-world manipulation settings show improved dynamics consistency and consistently higher task success.

LLM serving frameworks are quickly evolving with a complex software stack and a vast number of optimizations. The rapid development process can introduce silent errors where output quality silently degrades without any explicit error signals. Diagnosing silent errors is notoriously difficult due to the substantial semantic gap between the high-level symptoms and the low-level root causes. We observe that diagnosis of silent errors can be effectively framed as a differential debugging problem by leveraging the existence of semantically correct reference implementations. We propose Ekka, an automated diagnosis system that identifies root causes by systematically aligning and comparing intermediate execution states between a target and a reference framework. We constructed a benchmark of real-world silent errors from popular serving frameworks, where Ekka shows 80\% pass@$1$ diagnosis accuracy and 88\% pass@$5$ diagnosis accuracy, outperforming state-of-the-art systems. Ekka also diagnoses 4 new silent errors from serving frameworks, all of which have been confirmed by the developers.

Automated methods for discovering mechanistic simulator models from observational data offer a promising path toward accelerating scientific progress. Such methods often take the form of agentic-style iterative workflows that repeatedly propose and revise candidate models by imitating human discovery processes. However, existing LLM-based approaches typically implement such workflows via hand-crafted heuristic procedures, without an explicit probabilistic formulation. We recast model discovery as probabilistic inference, i.e., as sampling from an unknown distribution over mechanistic models capable of explaining the data. This perspective provides a unified way to reason about model proposal, refinement, and selection within a single inference framework. As a concrete instantiation of this view, we introduce ModelSMC, an algorithm based on Sequential Monte Carlo sampling. ModelSMC represents candidate models as particles which are iteratively proposed and refined by an LLM, and weighted using likelihood-based criteria. Experiments on real-world scientific systems illustrate that this formulation discovers models with interpretable mechanisms and improves posterior predictive checks. More broadly, this perspective provides a probabilistic lens for understanding and developing LLM-based approaches to model discovery.

We propose ***Var**iance **D**riven **E**xploration* (VarDE), a principled approach for pure exploration in *highly stochastic environments*, where the exploration process is dominated by stochastic variance. VarDE is built on a fundamental principle: *sampling effort should be allocated to minimize the uncertainty of the final decision*. We formalize the uncertainty of the final decision through a smooth decision function and derive allocation rules that explicitly capture how stochastic noise in individual components affects the reliability of the final output. We apply this methodology to three core problems of pure exploration -- Best Arm Identification (BAI), Monte Carlo Tree Search (MCTS), and Best-Policy Identification (BPI) -- with theoretical guarantees on variance decay and simple regret. Empirically, we demonstrate consistent and significant improvements of VarDE over existing methods, with especially strong gains in highly stochastic environments.

While Vision-Language-Action (VLA) models offer broad general capabilities, deploying them on specific hardware requires real-world adaptation to bridge the embodiment gap. Since robot demonstrations are costly, this adaptation must often occur under a strict data budget. In this work, we identify a critical **diversity trap**: the standard heuristic of ``maximizing coverage" by collecting diverse, single-shot demonstrations can be self-defeating due to non-vanishing estimation noise. We formalize this phenomenon as a **Coverage--Density Trade-off**. By decomposing the policy error into estimation (density) and extrapolation (coverage) terms, we characterize an interior optimal allocation of unique conditions for a fixed budget. Guided by this analysis, we propose **Anchor-Centric Adaptation (ACA)**, a two-stage framework that first stabilizes a policy skeleton through repeated demonstrations at core anchors, then selectively expands coverage to high-risk boundaries via teacher-forced error mining and constrained residual updates. Real-robot experiments validate our trade-off framework and demonstrate that ACA significantly improves task reliability and success rates over standard diverse sampling strategies under the same budget.

Graph foundation models face several fundamental challenges including transferability across diverse domains and data scarcity, which calls into question the very feasibility of creating such models. However, despite similar challenges, the tabular domain has recently witnessed the emergence of the first successful foundation models such as TabPFN. These models are based on the prior-data fitted networks (PFN) framework, in which models are pretrained on carefully designed synthetic datasets to make predictions in an in-context learning setting. Recently, G2T-FM, a framework that converts graph node-level tasks into tabular tasks, has made the first step towards adopting PFNs for graphs, yet it is limited to hand-crafted features and was never pretrained on graph data. In this work, we make the next step by proposing GraphPFN, a PFN-based model designed and pretrained specifically for graph node-level tasks. Following the PFN framework, we first design a prior distribution of synthetic attributed graphs by using a novel combination of multi-level stochastic block models and a preferential attachment process for structure generation and graph-aware structured causal models for attribute generation. Then, we augment the tabular foundation model LimiX with attention-based graph neighborhood aggregation layers and train it on millions of synthetic graphs sampled from our prior. On diverse real-world graph datasets with node-level tasks, GraphPFN achieves state-of-the-art results in both in-context learning and finetuning regimes, outperforming G2T-FM, prior GFMs, and task-specific GNNs trained from scratch. More broadly, GraphPFN shows the potential of PFN-based models for building graph foundation models. Our code is available at https://github.com/yandex-research/graphpfn.

In this work, we show that Latent Flow-Matching (LFM) models are robust to different types of perturbations, including data reduction and model capacity shrinkage. We characterize this stability by these models' tendency to generate similar outputs under identical noise seeds. We provide a perspective relating this phenomenon to flow matching theory, which indicates that this stability is inherent to the FM objective. We further exploit this stability to derive practical algorithms for more efficient training and inference. Concretely, first, we show that by training LFM models on significantly reduced datasets, performance is preserved, and in compute-constrained regimes, the model converges faster while maintaining quality. This yields multiple advantages, including savings in the training time due to faster convergence, and alleviating annotation effort when training conditional models. Second, LFM stability under architectural shrinkage gives rise to a two-model coarse-to-fine approach, one using a light-weight architecture for the first phase of the FM trajectory, and one with higher capacity for the second, thereby reducing the inference cost substantially. To determine which samples are informative, we introduce three sample-scoring criteria and evaluate them under standard metrics for generative models. Our results are thoroughly evaluated on multiple datasets, demonstrating the practical advantage of this stability, including data savings and a more than two-fold inference speedup while generating comparable outputs.

We present MoST (Mixture of Speech and Text), a novel multimodal large language model that seamlessly integrates speech and text processing through our proposed Modality-Aware Mixture of Experts (MAMoE) architecture. While current multimodal models typically process diverse modality representations with identical parameters—disregarding their inherent representational differences, we introduce specialized routing pathways that direct tokens to modality-appropriate experts based on input type. MAMoE simultaneously enhances modality-specific learning and cross-modal understanding through two complementary components: modality-specific expert groups that capture domain-specific patterns and shared experts that facilitate information transfer between modalities. Building on this architecture, we develop an efficient transformation pipeline that adapts the pretrained MoE language model through strategic post-training on ASR and TTS datasets, followed by fine-tuning with a carefully curated speech-text instruction dataset. A key feature of this pipeline is that it relies exclusively on fully accessible, open-source datasets to achieve strong performance and data efficiency. Comprehensive evaluations across ASR, TTS, audio language modeling, and spoken question answering benchmarks show that MoST consistently outperforms existing models of comparable parameter counts. Our ablation studies confirm that the modality-specific routing mechanism and shared experts design significantly contribute to performance gains across all tested domains. To our knowledge, MoST represents the first fully open-source speech-text LLM built on a Mixture of Experts architecture.

Offline Reinforcement Learning from Human Feedback (RLHF) pipelines such as Direct Preference Optimization (DPO) train on a pre-collected preference dataset, which makes them vulnerable to preference poisoning attack. We study label flip attacks against log-linear DPO. We first illustrate that flipping one preference label induces a parameter-independent shift in the DPO gradient. Using this key property, we can then convert the targeted poisoning problem into a structured binary sparse approximation problem. To solve this problem, we develop two attack methods: Binary-Aware Lattice Attack (BAL-A) and Binary Matching Pursuit Attack (BMP-A). BAL-A embeds the binary flip selection problem into a binary-aware lattice and applies Lenstra-Lenstra-Lovász reduction and Babai's nearest plane algorithm; we provide sufficient conditions that enforce binary coefficients and recover the minimum-flip objective. BMP-A adapts binary matching pursuit to our non-normalized gradient dictionary and yields coherence-based recovery guarantees and robustness (impossibility) certificates for $K$-flip budgets. Experiments on synthetic dictionaries and the Stanford Human Preferences dataset validate the theory and highlight how dictionary geometry governs attack success.

We investigate the relation between end-to-end equivariance and layerwise equivariance in deep neural networks. We prove the following: For a network whose end-to-end function is equivariant with respect to group actions on the input and output spaces, there is a parameter choice yielding the same end-to-end function such that its layers are equivariant with respect to some group actions on the latent spaces. Our result assumes that the parameters of the model are identifiable in an appropriate sense. This identifiability property has been established in the literature for a large class of networks, to which our results apply immediately, while it is conjectural for others. The theory we develop is grounded in an abstract formalism, and is therefore architecture-agnostic. Overall, our results provide a mathematical explanation for the emergence of equivariant structures in the weights of neural networks during training -- a phenomenon that is consistently observed in practice.

In this paper, we provide a comprehensive convergence analysis for the Lion optimizer. First, we establish that the original Lion achieves a convergence rate of $\mathcal{O}(d^{1/2}T^{-1/4})$, where $d$ denotes the problem dimension and $T$ is the iteration number. To improve this rate, we propose a variance reduction variant of Lion, which attains an enhanced rate of $\mathcal{O}(d^{1/2}T^{-1/3})$ with the average smoothness assumption. Then, we extend our analysis to distributed settings. We demonstrate that the distributed Lion optimizer and its variance reduction counterpart achieve linear speedup with respect to the number of nodes $n$, yielding convergence rates of $\mathcal{O}(d^{1/2}(nT)^{-1/4})$ and $\mathcal{O}(d^{1/2}(nT)^{-1/3})$, respectively. Additionally, we investigate a communication-efficient distributed Lion variant that utilizes sign compression for bidirectional communication. By employing unbiased sign operations, this variant achieves a convergence rate of $\mathcal{O} \left( \max \{ \frac{d^{1/4}}{T^{1/4}}, \frac{d^{1/10}}{n^{1/5}T^{1/5}} \} \right)$, and its variance-reduced counterpart can further improve the rate to $\mathcal{O}\left( \frac{d^{1/4}}{T^{1/4}} \right)$. Finally, we conduct numerical experiments to validate the effectiveness of the proposed methods.

Asynchronous pipeline parallelism maximizes hardware utilization by eliminating the pipeline bubbles inherent in synchronous execution, offering a path toward efficient large-scale distributed training. However, this efficiency gain can be compromised by gradient staleness, where the immediate model updates with delayed gradients introduce noise into the optimization process. Crucially, we identify a critical, yet often overlooked, pathology: this delay scales linearly with pipeline depth, fundamentally undermining the very scalability that the method originally intends to provide. We trace this pathology to a specific property of the optimization landscape: the misalignment between the Hessian eigenbasis and the standard coordinate basis, which triggers oscillations in the update trajectories of coordinate-wise adaptive optimizers. We identify that these oscillations cause delayed updates to diverge from their true counterparts, invalidating their use for current iterations. This insight is formalized through theoretical analysis, including a convergence bound showing that basis misalignment amplifies the delay penalty, and substantiated with empirical evaluation. To address this, we propose basis rotation, a framework that rotates the optimizer's coordinate system to align with the Hessian eigenbasis, keeping delayed updates useful. We theoretically demonstrate that basis rotation minimizes basis misalignment, thereby counteracting the conditions that amplify delay penalties. Empirically, in training up to a 3B-parameter LLM, basis rotation reduces the required iterations by 81.7% compared to the best-performing asynchronous baseline.

We introduce ECSEL, an explainable classification method that learns formal expressions in the form of signomial equations, motivated by the observation that many symbolic regression benchmarks admit compact signomial structure. ECSEL directly constructs a structural, closed-form expression that serves as both a classifier and an explanation. On standard symbolic regression benchmarks, our method recovers a larger fraction of target equations than competing state-of-the-art approaches while requiring substantially less computation. Leveraging this efficiency, ECSEL achieves classification accuracy competitive with established machine learning models without sacrificing interpretability. Further, we show that ECSEL satisfies some desirable properties regarding global feature behaviour, decision-boundary analysis, and local feature attributions. Experiments on benchmark datasets and two real-world case studies i.e., e-commerce and fraud detection, demonstrate that the learned equations expose dataset biases, support counterfactual reasoning, and yield actionable insights.

Surface reconstruction with differentiable rendering has achieved impressive performance in recent years, yet the pervasive photometric ambiguities have strictly bottlenecked existing approaches. This paper presents AmbiSuR, a framework that explores an intrinsic solution upon Gaussian Splatting for the photometric ambiguity-robust surface reconstruction with high performance. Started by revisiting the foundation, our investigation uncovers two built-in primitive-wise ambiguities in representation, while revealing an intrinsic potential for ambiguity self-indication in Gaussian Splatting. Stemming from these, a photometric disambiguation is first introduced, constraining ill-posed geometry solution for definite surface formation. Then, we propose an ambiguity indication module that unleashes the self-indication potential to identify and further guide correcting underconstrained reconstructions. Extensive experiments demonstrate our superior performance in surface reconstruction compared to existing methods across various challenging scenarios, while excelling in broad compatibility. Our code will be made open-source upon acceptance.

Visuomotor policies aim to learn complex manipulation tasks from expert demonstrations. However, generating smooth and coherent trajectories remains challenging, as it requires balancing proximal precision with distal foresight. Existing approaches typically focus on optimizing intra-chunk action distributions, often neglecting the inter-chunk coherence. Consequently, inter-chunk discontinuities significantly impede the learning of coherent long-horizon actions. To overcome this limitation and achieve a synergetic balance between precision and foresight, we propose FocalPolicy, a foresight-aware visuomotor policy that combines Frequency-Optimized Chunking with Locally Anchored flow matching. We introduce a foresight composite objective that supervises time-domain alignment within the proximal actions while regularizing frequency-domain structure over multiple future action chunks to improve cross-chunk coherence. To efficiently learn complex action distributions, we design locally anchored sampling to enhance target signal propagation efficiency during consistency flow matching training. Extensive experiments demonstrate that FocalPolicy outperforms existing approaches and confirm the generalizability of our modules to other baselines. Project website: https://focalpolicy.github.io/