ICML 2026 论文列表

Federated learning (FL) enables collaborative training across distributed edge devices, but deploying lightweight models in dynamic edge environments remains challenging. Existing methods typically require retraining whenever device resource constraints change, resulting in excessive computational overhead. We propose DFMPQ, a dynamic federated mixed-precision quantization framework that enables retraining-free deployment at the edge. DFMPQ trains a weight-sharing mixed-precision supernet via FL, which jointly represents diverse bit-width configurations. After training, resource-aware quantized subnets can be derived on demand to satisfy heterogeneous and time-varying resource constraints without additional optimization. Optimizing such a supernet in federated settings is difficult due to optimization interference among heterogeneous bit-widths and the coupling of quantization noise with non-IID data. DFMPQ addresses these issues through semantic-aware training and aggregation mechanisms that stabilize supernet optimization. In addition, a sensitivity-guided greedy search strategy is adopted to efficiently identify suitable quantization configurations under given resource budgets. Extensive experiments on multiple datasets and network architectures demonstrate that DFMPQ achieves competitive accuracy with significantly reduced computational cost, enabling efficient deployment for dynamic edge computing environments.

In statistical applications it has become increasingly common to encounter data structures that live on non-linear spaces such as manifolds. For data living on such non-linear spaces geodesic regression emerged as a natural extension of linear regression where the response variable lives on a Riemannian manifold. The parameters of geodesic regression capture the relationship of sensitive data, and hence, one should consider the privacy protection practices of said parameters. We consider releasing Differentially Private (DP) parameters of geodesic regression via the K-Norm Gradient (KNG) mechanism for Riemannian manifolds. We derive theoretical bounds for the sensitivity of the parameters showing they are tied to their respective Jacobi fields and hence the curvature of the space. We demonstrate the efficacy of our methodology on the sphere, $S_2\subset\mathbb{R}^3$, the space of symmetric positive definite matrices, and Kendall's planar shape space. Our methodology is general to any Riemannian manifold, and thus it is suitable for data in domains such as medical imaging and computer vision.

Modern high-throughput biological datasets containing thousands of perturbations enable large-scale discovery of causal graphs that represent regulatory interactions between genes. Differentiable causal graphical models and regression-based methods have been developed to infer gene regulatory networks (GRNs) from interventional datasets. However, existing approaches fail to capture the non-linear dynamics of biological processes such as cellular differentiation. To address this limitation, we propose $\textit{PerturbODE}$, a novel framework that employs interpretable neural ordinary differential equations (neural ODEs) to model cell state trajectories under perturbations and derive the underlying causal GRN from the neural ODE parameters, enabling downstream simulation of unseen genetic interventions. The GRN is encoded via a single-hidden-layer feedforward network, implicitly grouping genes into interpretable co-regulated modules. We demonstrate PerturbODE's efficacy in GRN inference and extension to perturbation response prediction across both simulated and real overexpression datasets.

Recent large language models (LLMs) have demonstrated strong capabilities across challenging tasks, enabling their widespread adoption in agentic systems that interact with external tools. In such deployments, however, LLMs are typically trained with general-purpose tokenizers designed for broad language coverage, while their usage is dominated by narrow, structured tool-calling interactions. This training–deployment mismatch leads to inefficient tokenization, where repetitive structural patterns and frequent semantic units in function calls are fragmented into long sequences of low-level tokens, increasing decoding overhead. To address this gap, we introduce $\textbf{AgentVocab}$, a structure-aware vocabulary adaptation framework for efficient LLM agents. AgentVocab derives specialized vocabulary entries from real tool-calling traces and adapts the model vocabulary to better reflect structural and semantic regularities, without task-specific schema engineering. Experiments on $\tau$-bench and $\tau^2$-bench show that AgentVocab preserves tool-calling performance while reducing latency relative to the vanilla baseline by 17.7\% and 19.5\%, respectively. Our approach is orthogonal to existing fine-tuning and agent-training methods and integrates seamlessly into standard agent pipelines. Source code and models will be available at https://github.com/Starry-159/AgentVocab.

Reward feedback learning (ReFL) is effective for both text-to-image (T2I) and text-to-video (T2V) generation with image reward models (RMs). However, image RMs are misaligned with temporal objectives of T2V, motivating ReFL with video reward models. Nevertheless, directly deploying video RMs is impractical due to their large parameter size and the prohibitive cost of memory. To address this, we propose VELR: an efficient framework that employs ensemble latent reward models (LRMs) to predict rewards directly in latent space, bypassing expensive backpropagation through VAE decoders and video RMs. Specifically, we introduce the ensemble technique for the LRM, which enhances capacity, quantifies uncertainty, and mitigates reward hacking. VELR achieves a reduction of up to 150GB in memory, requiring as little as 12.4% of the memory compared to standard ReFL. Experiments on OpenSora-1.2, CogVideoX-1.5, and Wan-2.1 with large-scale video RMs demonstrate that VELR achieves comparable performance as standard ReFL and enables efficient and robust video RM-based ReFL at scales previously unattainable.

Vision-language-action (VLA) models have advanced the field of embodied manipulation by harnessing broad world knowledge and strong generalization. However, current VLA models still face several key challenges, including limited reasoning capability, lack of status monitoring, and difficulty in self-correction. In this paper, we introduce \textbf{Sentinel-VLA}, a metacognitive VLA model equipped with an active ``sentinel'' module to monitor real-time execution status. Only when necessary, such as during initial planning or upon detecting an error, the model triggers a dynamic reasoning or formulate error recovery solutions. This on-demand reasoning mechanism ensures robust decision-making while minimizing computational overhead. Notably, all training data (spanning 44 tasks and over 2.6 million transitions) is automatically generated and annotated through our designed pipeline. We also propose the Self-Evolving Continual Learning (SECL) algorithm, which allows Sentinel-VLA to identify its capability boundaries and automatically collect data for expansion, paired with Orthogonal Continual Adapter (OC-Adapter) to constrain parameter updates to an orthogonal space, thereby preventing catastrophic forgetting. Real-world experiments demonstrate that Sentinel-VLA boosts the task success rate by over 30\% compared to the SOTA model, PI0. We will open-source all the code, weights, and data generation pipeline.

With the evolution of generative models, deepfakes have achieved near-perfect semantic realism, leaving forensic traces only in subtle structural anomalies. However, existing single-view paradigms often fail to generalize, as dominant semantic features overwhelm subtle artifact cues within entangled representations. This imbalance leads to overconfident yet brittle predictions—a phenomenon we term the Semantic Masking Effect. To address this challenge, we propose a reliable framework called Divide-and-Conquer Multi-View Evidential Learning (DiCoME) for Deepfake Detection. In the "Divide'' phase, we employ Geometric View Purification to decompose the entangled representation space through principled geometric projection. This process suppresses semantic interference within artifact-sensitive representations, forming the foundation for decorrelated yet complementary semantic and artifact views. In the "Conquer'' phase, we leverage Uncertainty-Aware Evidential Learning to synthesize these distinct views. By explicitly modeling the "epistemic conflict'' between semantic and artifact cues, this mechanism provides calibrated uncertainty estimates instead of forcing rigid deterministic decisions. Extensive experiments across multiple benchmarks demonstrate that our method consistently outperforms existing approaches in generalization performance, while providing reliable uncertainty estimation for trustworthy deepfake detection. Code is available at https://github.com/kxl0825/DiCoME.git.

Modern language models reason within bounded context, an inherent constraint that poses a fundamental barrier to long-horizon reasoning. We identify recursion as a core principle for overcoming this barrier, and propose recursive models as a minimal realization, where the model can recursively invoke itself to solve subtasks in isolated contexts. We prove that any computable problem admits a recursive decomposition of reasoning in which each subtask requires only exponentially smaller active context than standard autoregressive models; this strictly surpasses any context management approach confined to a single sequence, such as summarization. We further generalize our framework to modern agentic systems with arbitrary context processing and control flows, and prove that recursive models can achieve optimal power within this broader class. Experimentally, we train a 3B model to reason recursively and evaluate on Boolean satisfiability, a task requiring long-horizon combinatorial search, where it significantly outperforms frontier LLMs.

Accurate Zeroth-Order (ZO) Hessian estimation is a cornerstone of derivative-free methods, essential for tasks such as bilevel optimization, Bayesian inference, and uncertainty quantification. However, obtaining a complete suite of low-variance estimators for the Hessian and its inverse in high-dimensional settings remains a significant challenge. To address this, we propose a unified framework that reinterprets ZO Hessian approximation through the lens of single-step Policy Optimization (PO). This perspective establishes a theoretical equivalence between general ZO Hessian estimators and the Hessian of a smoothed PO objective, unifying distinct classical randomized estimators as specific instances of baseline selection. Building on this foundation, we introduce ZoVH, a comprehensive suite of variance-reduced estimators for the full Hessian matrix, its regularized inverse, and the bias-corrected inverse Hessian-gradient product. ZoVH leverages two key techniques: (1) a unique optimal baseline derived to provably minimize variance, and (2) a query reuse strategy that incorporates historical function queries to enhance sample efficiency without inflating costs. Our rigorous theoretical analysis confirms the unbiasedness of the Hessian estimator, validates the variance optimality of our baseline, provides error bounds for the entire ZoVH suite, and establishes convergence guarantees for the resulting curvature-aware ZO algorithm. Extensive empirical results validate our theoretical findings, demonstrating that ZoVH achieves superior estimation accuracy and convergence performance in real-world applications.

Dominant approaches for modelling Partial Differential Equations (PDEs) rely on deterministic predictions, yet many physical systems of interest are inherently chaotic and uncertain. While training probabilistic models from scratch is possible, it is computationally expensive and fails to leverage the significant resources already invested in high-performing deterministic backbones. In this work, we adopt a training-efficient strategy to transform pre-trained deterministic models into probabilistic ones via retrofitting with a proper scoring rule: the Continuous Ranked Probability Score (CRPS). Crucially, this approach is architecture-agnostic: it applies the same adaptation mechanism across distinct model backbones with minimal code modifications. The method proves highly effective across different scales of pre-training: for models trained on single dynamical systems, we achieve $20–54\%$ reductions in rollout CRPS and up to $30\%$ improvements in variance-normalised RMSE (VRMSE) relative to compute-matched deterministic fine-tuning. We further validate our approach on a PDE foundation model, trained on multiple systems and retrofitted on the dataset of interest, to show that our probabilistic adaptation yields an improvement of up to $40\%$ in CRPS and up to $15\%$ in VRMSE compared to deterministic fine-tuning. Validated across diverse architectures and dynamics, our results show that probabilistic PDE modelling need not require retraining from scratch, but can be unlocked from existing deterministic backbones with modest additional training cost.

Dueling optimization considers optimizing an objective with access to only a comparison oracle of the objective function. It finds important applications in emerging fields such as recommendation systems and robotics. Existing works on dueling optimization mainly focused on unconstrained problems in the Euclidean space. In this work, we study dueling optimization over Riemannian manifolds, which covers important applications that cannot be solved by existing dueling optimization algorithms. In particular, we propose a Riemannian Dueling Normalized Gradient Descent (RDNGD) method and establish its iteration complexity when the objective function is geodesically $L$-smooth or geodesically (strongly) convex. We also propose a projection-free algorithm, named Riemannian Dueling Frank–Wolfe (RDFW) method, to deal with the situation where projection is prohibited. We establish the iteration and oracle complexities for RDFW. We illustrate the effectiveness of the proposed algorithms through numerical experiments on both synthetic and real applications.

Parallelizing code for **irregular data structures** (sparse graphs, unbalanced trees, non-uniform meshes) is notoriously hard, and current LLMs fail catastrophically on such tasks, generating code riddled with race conditions, deadlocks, and poor scaling. We address this with **ParEVO**, a framework for synthesizing high-performance parallel algorithms for irregular data, built on three contributions: the **Parlay-Instruct Corpus** of 13,820 tasks generated via a "Critic-Refine" pipeline that filters for empirically performant uses of Work-Span primitives; specialized **DeepSeek**, **Qwen**, and **Gemini** models fine-tuned to the semantics of the ParlayLib library; and an **Evolutionary Coding Agent (ECA)** that repairs the "last mile" of correctness using compiler and profiler feedback. On the ParEval benchmark, ParEVO achieves an average **107x speedup** and a **13.6x speedup** on highly complex irregular graph problems, outperforming commercial models like GPT-5-Thinking and Gemini-3-Pro, while matching expert *human-written* baselines and reaching up to a **4.1x speedup** on kernels such as Maximal Independent Set. This demonstrates that AI-driven agents can effectively navigate the complex landscape of high-performance computing. Source code and datasets are available at https://github.com/WildAlg/ParEVO.

Leveraging future observation modeling to facilitate action generation presents a promising avenue for enhancing the capabilities of Vision-Language-Action (VLA) models. However, existing approaches struggle to strike a balance between maintaining efficient, predictable future representations and preserving sufficient fine-grained information to guide precise action generation. To address this limitation, we propose WoG (World Guidance), a framework that maps future observations into compact conditions by injecting them into the action inference pipeline. The VLA is then trained to simultaneously predict these compressed conditions alongside future actions, thereby achieving effective world modeling within the condition space for action inference. We demonstrate that modeling and predicting this condition space not only facilitates fine-grained action generation but also exhibits superior generalization capabilities. Moreover, it learns effectively from substantial human manipulation videos. Extensive experiments across both simulation and real-world environments validate that WoG significantly outperforms existing methods based on future prediction. Project page is available at: https://selen-suyue.github.io/WoGNet/.

While recent autoregressive models have achieved text-to-image generation performance comparable to diffusion models, they significantly struggle with fine-grained semantic alignment. To rigorously evaluate this limitation, we introduce DeltaBench, a benchmark featuring paired prompts with subtle fine-grained differences, which reveals that existing models fail to achieve precise control over visual tokens. To bridge this gap, we propose FocusDiff, a comprehensive framework that enhances alignment by learning from subtle differences in similar text-image pairs. Specifically, we construct FocusDiff-Data, a large-scale dataset of paired samples derived from image editing tasks to capture localized semantic shifts. Furthermore, we introduce Pair-GRPO, an improved reinforcement learning algorithm that extends GRPO to paired samples. Extensive experiments demonstrate that our approach outperforms most prior prominent methods on both DeltaBench and existing benchmarks.

Recent work describes what transformers can and cannot compute through connections to boolean circuits, but existing results lack exact characterizations and are sensitive to modeling choices. Padded transformers---to whose input filler symbols such as ``...'' are appended---emerge as a useful gadget for establishing equivalences to circuit classes by providing polynomial space for adaptive parallel computation. However, only a limited set of padded transformer idealizations has been studied, leaving open how robustly these equivalences hold under changes to attention type, model width, and uniformity. We find that, under practical assumptions, padded transformers are surprisingly robust to all of these, and identify numeric precision and model depth as the main factors affecting expressivity. Concretely, we prove that polynomially padded $\text{L-uniform}$ constant-precision transformers are equivalent to $\text{L-uniform AC}^0$, while growing-precision ones achieve $\text{L-uniform TC}^0$ regardless of width. Furthermore, looping enables sequential processing analogous to circuits: $\log^d N$-looped constant-precision transformers reach $\text{FO-uniform AC}^d$, and growing-precision ones reach $\text{FO-uniform TC}^d$. Interestingly, growing width or precision beyond logarithmic does not increase expressivity, and all our results hold for both softmax and average hard attention transformers.

Partial Differential Equations (PDEs) play a fundamental role in scientific computing, and recent efforts have sought to extend the success of foundation models to PDE solving. However, multi-physics PDE pre-training faces the unique challenge of disentangling dynamic heterogeneity to learn universal, elementary patterns that generalize to new PDEs. Additionally, cross-physics transfer lacks a theoretical framework for interpretability—specifically, understanding which pre-trained operator knowledge is effectively transferred to target PDEs. To bridge these gaps, we introduce the theory of neural operator splitting, which decomposes PDE evolution into a modulated global spectral operator and sparse local constitutive mechanisms. A key innovation is Origo, which provides a neural operator bank that enables the identification of operator-level generalization patterns. Extensive experiments demonstrate strong zero-shot generalization and mechanism-level interpretability on unseen PDEs.

Representing the past in a compressed, efficient, and informative manner is a central problem for systems trained on sequential data. The \textit{HiPPO} framework, originally proposed by Gu \& Dao et al., provides a principled approach to sequential compression by projecting signals onto orthogonal polynomial (OP) bases via structured linear ordinary differential equations. Subsequent works have embedded these dynamics in state space models (SSMs), where HiPPO structure serves as an initialization. Nonlinear successors of these SSM methods such as Mamba are state-of-the-art for many tasks with long-range dependencies, but the mechanisms by which they represent and prioritize history remain largely implicit. In this work, we revisit the HiPPO framework with the goal of making these mechanisms explicit. We show how polynomial representations of history can be extended to support capabilities of modern SSMs such as adaptive memory allocation and associative memory, while retaining direct interpretability in the OP basis. We introduce a unified framework comprising five such extensions, which we collectively refer to as a ``HiPPO zoo.'' Each extension exposes a specific modeling capability through an explicit, interpretable modification of the HiPPO framework. The resulting models adapt their memory online and train in streaming settings with efficient updates. We illustrate the behaviors and modeling advantages of these extensions through a range of synthetic sequence modeling tasks, demonstrating that capabilities typically associated with modern SSMs can be realized through explicit, interpretable polynomial memory structures.

Finding mathematical relations underlying natural phenomena is a fundamental task in scientific discovery. Recent advances in evolutionary search with Large Language Models (LLMs) show great promise by leveraging their embedded scientific knowledge. However, discovering governing equations remains challenging due to vast combinatorial hypothesis spaces with exponentially many possible relations. Existing LLM-based approaches treat LLMs as static hypothesis generators unaware of the observed scientific system, leading to suboptimal and inefficient exploration that over-relies on internal priors. To address this, we introduce Decompose, Adapt, and Evolve (DecAEvolve), a framework that combines granular feedback from symbolic term decomposition with LLM refinement through reinforcement learning fine-tuning. DecAEvolve unifies symbolic decomposition with test-time RL adaptation, enabling adaptive rather than static hypothesis generation. Our experiments across diverse scientific benchmarks demonstrate that DecAEvolve significantly improves both the accuracy of discovered equations and the efficiency of the discovery process, reducing error by up to an order of magnitude compared to state-of-the-art baselines.

Reinforcement learning with verifiable rewards (RLVR) is a cornerstone of post-training for large reasoning models, yet widely used algorithms such as Group Relative Policy Optimization (GRPO) often exhibit \textbf{diversity collapse}. We provide a geometric diagnosis by formalizing GRPO as a dynamical flow on the probability simplex. Under a mode-based coarse-graining of rollouts, we show that GRPO induces a \textbf{collision field} over correct modes, monotonically pushing towards simplex vertices and thus yielding a \textbf{winner-take-all} regime. To address this systematically, we introduce \textbf{G$^2$RPO (Geometric GRPO)}, which reshapes RLVR via principled \textbf{vector-field editing}. Concretely, we intervene at the advantage level by adding granularity bonuses inversely proportional to mode probabilities, encouraging underrepresented correct modes. The bonus has a natural geometric interpretation, and its potential performance side effects can be mitigated, thereby avoiding the usual accuracy--diversity trade-off. In experiments with 7B and 14B models trained on a math reasoning task and evaluated on \textbf{AIME 2024/2025}, GRPO loses up to \textbf{57\%} of active correct modes. In contrast, G$^{2}$RPO increases active correct-mode coverage by \textbf{172\%--205\%}, reduces concentration on any single correct mode, prevents the late-stage \emph{entropy crash}, and improves \texttt{pass@1} by \textbf{+1.4} to \textbf{+7.9} points relative to GRPO. Overall, diversity is not merely a regularizer but a \textbf{geometric property} to be controlled to improve the model without trapping it in a single dominant strategy.