ICML 2026 Papers

In this paper, we study the private and robust alignment of language models from a theoretical perspective by establishing upper bounds on the suboptimality gap in both offline and online settings. We consider preference labels subject to privacy constraints and/or adversarial corruption, and analyze two distinct interplays between them: privacy-first and corruption-first. For the privacy-only setting, we show that log loss with an MLE-style algorithm achieves near-optimal rates, in contrast to conventional wisdom. For the joint privacy-and-corruption setting, we first demonstrate that existing offline algorithms in fact provide stronger guarantees---simultaneously in terms of corruption level and privacy parameters---than previously known, which further yields improved bounds in the corruption-only regime. In addition, we also present the first set of results for private and robust online alignment. Our results are enabled by new uniform convergence guarantees for log loss and square loss under privacy and corruption, which we believe have broad applicability across learning theory and statistics.

The stochastic minimax optimization problem has widespread applications in machine learning. Recently, numerous distributed minimax optimization algorithms have been developed to handle distributed training data. However, most of these algorithms suffer from high communication costs. To address this issue, we develop a novel communication-efficient decentralized stochastic gradient descent ascent with momentum algorithm based on the error feedback mechanism. Importantly, our algorithm demonstrates how to balance the full-precision update and the compression residual with novel designs for coefficients regarding variables and gradients to guarantee convergence. However, compressing the primal and dual variables (and their gradients) of stochastic minimax optimization problems with the error feedback mechanism presents significant challenges for convergence analysis. In particular, it incurs the circle dependence among consensus errors and compression errors. To overcome this challenge, we propose novel strategies that enable the establishment of the convergence rate for our algorithm. Our theoretical results demonstrate how the compression operator influences the convergence rate. Finally, extensive experimental results confirm the efficacy of our proposed algorithm.

Compositional diffusion planners aim to solve long-horizon robotic tasks using short training trajectories. Yet, current approaches often rely on the heuristic stitching of local predictions. We show that the resulting stitched update is generally a non-conservative field} that does not mathematically correspond to any valid global trajectory log-density function. We propose **Energy-based Compositional Diffuser** (ECD), a framework that formulates the global trajectory as the minimizer of the sum of local bridge potentials. This energy-based perspective defines a conservative correction field and contains a boundary reaction term that heuristic stitching omits. To enable efficient inference, we further introduce a Markov-based score approximation that computes the reaction term via a single block-tridiagonal solve, maintaining time complexity linear in the planning horizon. Empirically, ECD achieves state-of-the-art success rates on a range of OGBench stitching tasks, while nearly matching the inference speed of heuristic stitching methods. Code is available at https://github.com/GradientSpaces/ECD.

While large vision-language models (VLMs) show promise for object goal navigation, current methods still struggle with low success rates and inefficient localization of unseen objects—failures primarily attributed to weak temporal-spatial reasoning. Meanwhile, recent attempts to inject reasoning into VLM-based agents improve success rates but incur substantial computational overhead. To address both the ineffectiveness and inefficiency of existing approaches, we introduce Hydra-Nav, a unified VLM architecture that adaptively switches between a deliberative "slow system" for analyzing exploration history and formulating high-level plans, and a reactive "fast system" for efficient execution. We train Hydra-Nav through a three-stage curriculum: (i) spatial-action alignment to strengthen trajectory planning, (ii) memory-reasoning integration to enhance temporal-spatial reasoning over long-horizon exploration, and (iii) iterative rejection fine-tuning to enable selective reasoning at critical decision points. Extensive experiments demonstrate that Hydra-Nav achieves state-of-the-art performance on the HM3D, MP3D, and OVON benchmarks, outperforming the second-best methods by 11.1\%, 17.4\%, and 21.2\%, respectively. Furthermore, we introduce SOT (Success weighted by Operation Time), a new metric to measure search efficiency across VLMs with varying reasoning intensity. Results show that adaptive reasoning significantly enhances search efficiency over fixed-frequency baselines. Project page is available at: https://zixuan-wang99.github.io/Hydra-Nav/.

Large generative models raise growing concerns about provenance, misinformation, and impersonation. Digital watermarking offers a principled solution, yet extending it to natural language remains challenging due to text discreteness and sensitivity to semantic perturbations. Existing text watermarking methods either operate at the token level requiring white-box access and remaining fragile to paraphrasing, or at the sentence level, which supports black-box deployment but suffers from low Watermark Success Rate (WSR). We show that low WSR in sentence-level watermarking primarily stems from low injection success probability caused by a mismatch between posterior embedding distributions and semantic accept regions. Based on this insight, we propose \textbf{X-Guard}, a geometry-aware sentence-level watermarking framework that improves injection success by jointly optimizing embedding distributions and semantic space partitioning. X-Guard learns a more isotropic embedding space and introduces \textbf{A$^2$PQ}, a centroid-aligned partitioning scheme that approximately equalizes probability mass across regions. Extensive experiments across multiple models, languages, and attack settings demonstrate that X-Guard consistently improves robustness while preserving text fluency and practical deployability.

Latent actions learned from diverse human videos serve as pseudo-labels for vision-language-action (VLA) pretraining, but provide effective supervision only if they remain informative about the underlying ground-truth actions. For effective supervision, latent actions should contain information about the underlying actions even though they are inaccessible. We propose **M**ulti-**V**iew**P**oint **L**atent **A**ction **M**odel (**MVP-LAM**), which learns latent actions that are highly informative about ground-truth actions from multi-view videos. MVP-LAM trains latent actions with a *cross-viewpoint reconstruction* objective, so that a latent action from one view must explain the future in another view, reducing reliance on viewpoint-specific cues. On Bridge V2, MVP-LAM produces more action-centric latent actions, achieving higher mutual information with ground-truth actions and improved action prediction, including under out-of-distribution evaluation. Finally, pretraining VLAs with MVP-LAM latent actions improves downstream manipulation performance on various benchmarks. The code and trained checkpoints are available at [https://jm-this.github.io/mvp_lam/](https://jm-this.github.io/mvp_lam/).

Current vision language models face hallucination and robustness issues against ambiguous or corrupted modalities. We hypothesize that these issues can be addressed by exploiting the shared information between modalities to compensate for the impaired one. To this end, we analyze multimodal interactions -- redundant (shared), unique (exclusive), and synergistic (emergent) task-relevant information provided by the modalities -- to determine their impacts on model reliability. Specifically, amplifying redundant interactions would increase this exploitable shared information to resolve these issues; yet, modern instruction datasets often eliminate redundancies to prioritize visual grounding. We bridge this gap through a self-captioning workflow featuring a \textsc{Multimodal Interaction Gate}: a mechanism to convert unique interactions into redundant interactions. Our findings suggest that increasing redundancy can reduce visual induced errors by 38.3\% and improve consistency by 16.8\%.

Mean Field Games (MFGs) provide a powerful framework for modeling the collective behavior of large populations of interacting agents. In this paper, we address the problem of Imitation Learning (IL) in MFGs subject to common noise, where the population distribution evolves stochastically. This stochasticity compels agents to adopt population-aware policies to respond to aggregate shocks. We formulate two distinct learning objectives: recovering a Nash equilibrium and maximizing performance against an expert population. We investigate two imitation proxies: Behavioral Cloning (BC) and Adversarial (ADV) divergence. We then establish finite-sample error bounds showing that minimizing these proxies effectively controls both the policy’s exploitability and its performance gap relative to the expert. Furthermore, we propose a numerical framework using generalized Fictitious Play and Deep Learning to compute expert population-aware policies. Through experiments on three environments we demonstrate that standard population-unaware policies fail to capture the equilibrium dynamics. Our results highlight that learning population-aware policies is crucial to avoid being misled by the randomness inherent in common noise.

Fine-grained cross-modal alignment is pivotal for multimodal reasoning yet remains limited by Semantic Sparsity Bias—a fundamental asymmetry where dense visual signals are under-represented by sparse textual captions. This disparity leads to the inadvertent suppression of contextually vital visual regions (patch redundancy) and hinders precise concept grounding (patch ambiguity). While Multimodal Large Language Models (MLLMs) offer rich descriptive capabilities, their naive integration often induces semantic drift due to inconsistencies with sparse ground-truth captions. To systematically resolve these challenges, we present the Semantic-Enhanced Patch Slimming (SEPS) framework. Central to SEPS is a novel Dual-Granularity Semantic Calibration mechanism, which synthesizes a Holistic Visual-Linguistic Anchor from MLLMs to synergize with original sparse queries. This mechanism transforms patch selection into a semantic consensus process, ensuring that retained patches satisfy both local discriminability and global contextual integrity. Furthermore, we propose a Salience-Guided Metric Aggregation strategy to mitigate the similarity dilution effect inherent in global mean pooling, thereby amplifying highly-relevant patch-word correspondences. Extensive experiments on Flickr30K and MS-COCO datasets demonstrate that SEPS surpasses existing state-of-the-art approaches across diverse backbones, delivering significant performance gains in text-to-image retrieval tasks. The complete implementation is available at https://anonymous.4open.science/r/SEPS/.

Online reinforcement learning in non-episodic, finite-horizon MDPs remains underexplored and is challenged by the need to estimate returns to a fixed terminal time. Existing infinite-horizon methods, which often rely on discounted contraction, do not naturally account for this fixed-horizon structure. We introduce a modified Q-function: rather than targeting the full-horizon, we learn a K-step lookahead Q-function that truncates planning to the next K steps. To further improve sample efficiency, we introduce a thresholding mechanism: actions are selected only when their estimated K-step lookahead value exceeds a time-varying threshold. We provide an efficient tabular learning algorithm for this novel objective, proving it achieves fast finite-sample convergence: it achieves minimax optimal constant regret for $K=1$ and $\mathcal{O}(\max((K-1),C_{K-1})\sqrt{SAT\log(T)})$ regret for any $K \geq 2$. We numerically evaluate the performance of our algorithm under the objective of maximizing reward. Our implementation adaptively increases K over time, balancing lookahead depth against estimation variance. Empirical results demonstrate superior cumulative rewards over state-of-the-art tabular RL methods across synthetic MDPs and RL environments: JumpRiverswim, FrozenLake and AnyTrading. Code is provided on \href{https://github.com/jamie01713/K-Step-Lookahead}{github}.

Knowledge Graph Question Answering (KGQA) offers grounded, interpretable reasoning, but existing methods often fail to provide reliable coverage guarantees over retrieved answers. While Conformal Prediction (CP) offers a principled framework for producing prediction sets with statistical guarantees, prior conformal KGQA methods suffer from two critical pitfalls: violated coverage guarantees due to invalid calibration, and weak score discriminability that yields excessively large prediction sets. We propose Conformal Path Reasoning (CPR), a novel trustworthy KGQA framework built on two key innovations. First, query-level conformal calibration over path-level scores preserves exchangeability to ensure valid coverage guarantees. Second, we introduce the Residual Conformal Value Network (RCVNet), a lightweight module trained via PUCT-guided exploration to learn discriminative path-level nonconformity scores. Extensive experiments show that CPR significantly improves the Empirical Coverage Rate by 45\% while reducing prediction set size by 52\% on average over conformal baselines across benchmark datasets, highlighting its effectiveness for reliable conformal reasoning over knowledge graphs.

The *de novo* generation of high-affinity epitopes tailored to specific major histocompatibility complex (MHC) proteins is a pivotal challenge in computational immunotherapy. However, current methods struggle to effectively integrate the MHC context into the generation process, and often fail to guarantee high binding affinity due to the neglect of discriminative signals from non-binders. To bridge these gaps, we present **EpiCoCo**, a probabilistic framework for **Epi**tope generation via MHC-context **Co**-modeling and **Co**ntrastive affinity learning. EpiCoCo treats the pMHC complex as a dynamic, co-adaptive system by operating on the joint E(3) graph. In addition, we introduce Contrastive Affinity Guidance (CAG), an inference mechanism that leverages the gradient difference between learned high- and low-affinity distributions. CAG actively drives the generation trajectory towards high-affinity manifolds while utilizing repulsive signals to filter out candidates with poor binding potential. Extensive evaluations demonstrate that EpiCoCo achieves a mean binding free energy of -45.20 REU, a 23% improvement over the state-of-the-art, while maintaining high structural plausibility. The results validate that context co-modeling and negative-informed guidance are essential for generating valid, high-potency immunotherapeutics.

Large language models have advanced web agents, yet current agents lack personalization capabilities. Since users rarely specify every detail of their intent, practical web agents must be able to interpret ambiguous queries by inferring user preferences and contexts. To address this challenge, we present Persona2Web, the first benchmark for evaluating personalized web agents on the real open web, built upon the clarify-to-personalize principle, which requires agents to resolve ambiguity based on user history rather than relying on explicit instructions. Persona2Web consists of: (1) user histories that reveal preferences implicitly over long time spans, (2) ambiguous queries that require agents to infer implicit user preferences, and (3) a reasoning-aware evaluation framework that enables fine-grained assessment of personalization. We conduct extensive experiments across various agent architectures, backbone models, history access schemes, and queries with varying ambiguity levels, revealing key challenges in personalized web agent behavior. For reproducibility, our codes and datasets are publicly available at https://serin-kimm.github.io/Persona2Web/

Evaluation benchmarks are essential for assessing vision--language models (VLMs), but most multimodal benchmarks are static, making them vulnerable to temporal staleness, data contamination, and costly maintenance. We present MMBench-Live, a continuously evolving multimodal benchmark built by a multi-agent-driven automated pipeline. Our framework treats benchmark evolution as task-guided dataset construction, integrating structured benchmark specification, feedback-controlled real-time data acquisition, and verifiable QA generation with executable reasoning. To maintain cross-version comparability, we introduce a distribution-consistent update strategy that extracts task-related visual patterns from the original benchmark to guide data collection and filtering. Instantiated from MMBench, MMBench-Live contains 5.9K newly generated evaluation instances with a high answer correctness rate, while each update costs about \$30 and takes 1--2 hours. Extensive evaluations show that MMBench-Live preserves stable model rankings, maintains semantic alignment with the original benchmark, and exhibits weaker contamination-related memorization signals, suggesting a practical and scalable paradigm for sustainable multimodal benchmark evolution. The project is available at \url{https://github.com/PRIS-CV/MMBench-Live}.

Reinforcement learning significantly enhances LLM capabilities but suffers from a critical issue: length inflation, where models adopt verbosity or inefficient reasoning to maximize rewards. Prior approaches struggle to address this challenge in a general and lossless manner, primarily because additive penalties introduce a compensatory effect that creates optimization shortcuts, while heuristic gating strategies lack generality beyond binary feedback. To bridge this gap, we present Group Relative Reward Rescaling (GR$^3$), which reframes length control as a multiplicative rescaling paradigm, effectively establishing a generalized, continuous, and reward-dependent gating mechanism. To further ensure lossless optimization, we incorporate group-relative regularization and advantage-aware calibration, which dynamically adapt length budgets to instance difficulty and preserve the advantage signal of high-quality trajectories. Empirically, across both RLHF and RLVR settings, GR$^3$ maintains training dynamics and downstream performance comparable to standard GRPO while significantly mitigating length inflation, outperforming state-of-the-art length-regularized baselines.

LLM pre-training efficacy increasingly depends on data composition rather than sheer volume. Yet, optimal mixing is hindered by categorization flaws: human taxonomies suffer from ontological misalignment, and Euclidean clustering fails to address embedding anisotropy. We introduce **GEM** (**G**eometric **E**ntropy **M**ixing), a framework reformulating data curation as a variational problem on the hypersphere augmented with a **mixing-balance regularizer**. By decoupling the generative prior and optimizing the objective via a provable **MM (Minorize-Maximize)** algorithm, GEM effectively counteracts the cluster collapse to discover balanced semantic structures invisible to Euclidean heuristics. We employ teacher-student distillation to scale this geometric fidelity to web-scale corpora and introduce the **Geometric Influence Score (GIS)** for interpretable taxonomy generation. Experiments with 1.1B-parameter models demonstrate that GEM establishes a new state-of-the-art when integrated into mixing strategies like DoReMi and RegMix, improving average downstream accuracy by up to **1.2%** and offering a robust coordinate system for predictable data mixing.

\textit{\underline{M}ultiple \underline{k}ernel \underline{l}earning}~(MKL) combines several base kernels in the spirit of ensemble learning, yet existing methods rarely model kernel diversity---a known cornerstone of ensembles---and most traditional kernels weight all features uniformly, ignoring feature-level discriminability. We address both gaps with DARFMMKL: a data-driven kernel family (Recursive Feature Machine kernels) that learns feature importance directly from data, paired with a kernel selection method that jointly optimizes diversity and quality. The resulting NP-hard binary quadratic program is reformulated via Glover linearization and continuous relaxation into a linear program, and accelerated by Nystr\"om sketching, yielding a selector whose cost is decoupled from the sample size. We provide a covering-number generalization bound that explicitly relates kernel diversity to estimation error. Experiments on 12 benchmark datasets show that DARFMMKL consistently outperforms 9 state-of-the-art MKL methods.

Multi-modal spatio-temporal forecasting underpins many real-world applications but remains challenging due to the complex and evolving interactions across modalities and time steps. Moreover, the lack of interpretability in existing models limits their reliability in safety-critical scenarios. In this paper, we present CausalX, a unified and causally interpretable plug-and-play model for multi-modal spatio-temporal forecasting. CausalX achieves interpretability by learning a dynamic causal graph across modalities and time, whose edge weights quantify causal attribution strength, and are further refined by a diffusion-based generative process guided by structural priors. To overcome the absence of ground-truth causal structures, CausalX aggregates multi-source constraints from causal analysis techniques and a variational autoencoder, spanning predictive, temporal, interventional, and generative aspects to jointly learn a more comprehensive causal graph. Extensive experiments on real-world forecasting tasks, including pedestrian trajectory prediction and tropical cyclone forecasting, demonstrate that CausalX achieves superior accuracy while producing interpretable causal graphs. CausalX is modular, architecture-agnostic, and generalizable, offering a new perspective on bridging causal inference and spatio-temporal forecasting.

Multi-objective reinforcement learning (MORL) is a fundamental framework for real-world decision-making problems involving multiple conflicting criteria. Existing multi-policy (MP) methods typically rely on online evolutionary frameworks that maintain large policy populations, leading to high sample complexity and excessive agent–environment interactions. To mitigate these limitations, we present Multi-policy Pareto Front Tracking (MPFT), a framework without a self-evolving population. It leverages an efficient Pareto-tracking mechanism initialized with single-objective extreme policies to trace the Pareto front, and further densifies sparse regions to achieve an accurate approximation of the full Pareto front. MPFT can be seamlessly integrated with advanced offline MORL algorithms, thereby substantially improving sample efficiency. We evaluate MPFT on six robotic control tasks with up to three objectives and three high-dimensional tasks with more than three objectives. Experimental results show that MPFT outperforms state-ofthe-art baselines in terms of hypervolume and expected utility. It also significantly reduces agent–environment interactions. These results further demonstrate that MPFT serves as a general purpose framework that can seamlessly integrate both online and offline MORL algorithms.