ICML 2025 论文列表

The infringement risks of LLMs have raised significant copyright concerns across different stages of the model lifecycle. While current methods often address these issues separately, this position paper argues that the LLM copyright challenges are inherently connected, and independent optimization of these solutions leads to theoretical bottlenecks. Building on this insight, we further argue that managing LLM copyright risks requires a systemic approach rather than fragmented solutions. In this paper, we analyze the limitations of existing methods in detail and introduce an iterative online-offline joint optimization framework to effectively manage complex LLM copyright risks. We demonstrate that this framework offers a scalable and practical solution to mitigate LLM infringement risks, and also outline new research directions that emerge from this perspective.

Advances in large models, reinforcement learning, and open-endedness have accelerated progress toward autonomous agents that can learn and interact in the real world. To achieve this, flexible tools are needed to create rich, yet computationally efficient, environments. While scalable 2D environments fail to address key real-world challenges like 3D navigation and spatial reasoning, more complex 3D environments are computationally expensive and lack features like customizability and multi-agent support. This paper introduces Craftium, a highly customizable and easy-to-use platform for building rich 3D single- and multi-agent environments. We showcase environments of different complexity and nature: from single- and multi-agent tasks to vast worlds with many creatures and biomes, and customizable procedural task generators. Benchmarking shows that Craftium significantly reduces the computational cost of alternatives of similar richness, achieving +2K steps per second more than Minecraft-based frameworks.

Safe offline reinforcement learning (RL), which aims to learn the safety-guaranteed policy without risky online interaction with environments, has attracted growing recent attention for safety-critical scenarios. However, existing approaches encounter out-of-distribution problems during the testing phase, which can result in potentially unsafe outcomes. This issue arises due to the infinite possible combinations of reward-related and cost-related states. In this work, we propose State Decoupling with Q-supervised Contrastive representation (SDQC), a novel framework that decouples the global observations into reward- and cost-related representations for decision-making, thereby improving the generalization capability for unfamiliar global observations. Compared with the classical representation learning methods, which typically require model-based estimation (e.g., bisimulation), we theoretically prove that our Q-supervised method generates a coarser representation while preserving the optimal policy, resulting in improved generalization performance. Experiments on DSRL benchmark problems provide compelling evidence that SDQC surpasses other baseline algorithms, especially for its exceptional ability to achieve almost zero violations in more than half of the tasks, while the state-of-the-art algorithm can only achieve the same level of success in a quarter of the tasks. Further, we demonstrate that SDQC possesses superior generalization ability when confronted with unseen environments.

Large language models (LLMs) provide excellent text-generation capabilities, but standard prompting and generation methods generally do not lead to intentional or goal-directed agents and might necessitate considerable prompt tuning. Even the best current LLMs rarely ask clarifying questions, engage in explicit information gathering, or take actions that lead to better decisions after multiple turns. Reinforcement learning has the potential to leverage the powerful modeling capabilities of LLMs, as well as their internal representation of textual interactions, to create capable goal-directed language agents. This can enable intentional and temporally extended interactions, such as with humans, the emergence of complex skills such as persuasion, and long-horizon strategic behavior, such as in the context of games. Enabling this requires the community to develop reliable reinforcement learning algorithms for training LLMs. Developing such algorithms requires tasks that can gauge progress on algorithm design, provide accessible and reproducible evaluations for multi-turn interactions, and cover a range of task properties and challenges in improving reinforcement learning algorithms. Our paper introduces the LMRL-Gym benchmark for evaluating multi-turn RL for LLMs, together with an open-source research framework for getting started on multi-turn RL with offline value-based and online policy-based RL methods. Our benchmark consists of 3 Interactive Dialogue tasks and 5 RL Capability tests for a total of 8 tasks, which require multiple rounds of language interaction and cover a range of tasks in open-ended dialogue and text games.

Pre-trained generalist policies are rapidly gaining relevance in robot learning due to their promise of fast adaptation to novel, in-domain tasks. This adaptation often relies on collecting new demonstrations for a specific task of interest and applying imitation learning algorithms, such as behavioral cloning. However, as soon as several tasks need to be learned, we must decide *which tasks should be demonstrated and how often?* We study this multi-task problem and explore an interactive framework in which the agent *adaptively* selects the tasks to be demonstrated. We propose AMF (Active Multi-task Fine-tuning), an algorithm to maximize multi-task policy performance under a limited demonstration budget by collecting demonstrations yielding the largest information gain on the expert policy. We derive performance guarantees for AMF under regularity assumptions and demonstrate its empirical effectiveness to efficiently fine-tune neural policies in complex and high-dimensional environments.

Retrieval-augmented generation (RAG) systems face a fundamental challenge in aligning independently developed retrievers and large language models (LLMs). Existing approaches typically involve modifying either component or introducing simple intermediate modules, resulting in practical limitations and sub-optimal performance. Inspired by human search behavior—typically involving a back-and-forth process of proposing search queries and reviewing documents, we propose C-3PO, a proxy-centric framework that facilitates communication between retrievers and LLMs through a lightweight multi-agent system. Our framework implements three specialized agents that collaboratively optimize the entire RAG pipeline without altering the retriever and LLMs. These agents work together to assess the need for retrieval, generate effective queries, and select information suitable for the LLMs. To enable effective multi-agent coordination, we develop a tree-structured rollout approach for reward credit assignment in reinforcement learning. Extensive experiments in both in-domain and out-of-distribution scenarios demonstrate that C-3PO significantly enhances RAG performance while maintaining plug-and-play flexibility and superior generalization capabilities.

Diplomacy is a complex multiplayer game that re- quires both cooperation and competition, posing significant challenges for AI systems. Traditional methods rely on equilibrium search to generate extensive game data for training, which demands substantial computational resources. Large Lan- guage Models (LLMs) offer a promising alterna- tive, leveraging pre-trained knowledge to achieve strong performance with relatively small-scale fine-tuning. However, applying LLMs to Diplo- macy remains challenging due to the exponential growth of possible action combinations and the intricate strategic interactions among players. To address this challenge, we propose DipLLM, a fine-tuned LLM-based agent that learns equilib- rium policies for Diplomacy. DipLLM employs an autoregressive factorization framework to sim- plify the complex task of multi-unit action assign- ment into a sequence of unit-level decisions. By defining an equilibrium policy within this frame- work as the learning objective, we fine-tune the model using only 1.5% of the data required by the state-of-the-art Cicero model, surpassing its per- formance. Our results demonstrate the potential of fine-tuned LLMs for tackling complex strategic decision-making in multiplayer games.

Boosting provides a practical and provably effective framework for constructing accurate learning algorithms from inaccurate rules of thumb. It extends the promise of sample-efficient learning to settings where direct Empirical Risk Minimization (ERM) may not be implementable efficiently. In the realizable setting, boosting is known to offer this computational reprieve without compromising on sample efficiency. However, in the agnostic case, existing boosting algorithms fall short of achieving the optimal sample complexity. We highlight a previously unexplored avenue of improvement: unlabeled samples. We design a computationally efficient agnostic boosting algorithm that matches the sample complexity of ERM, given polynomially many additional unlabeled samples. In fact, we show that the total number of samples needed, unlabeled and labeled inclusive, is never more than that for the best known agnostic boosting algorithm -- so this result is never worse -- while only a vanishing fraction of these need to be labeled for the algorithm to succeed. This is particularly fortuitous for learning-theoretic applications of agnostic boosting, which often take place in the distribution-specific setting, where unlabeled samples can be availed for free. We also prove that the resultant guarantee is resilient against mismatch between the distributions governing the labeled and unlabeled samples. Finally, we detail an application of this result in reinforcement learning.

The integration of Deep Neural Networks in Reinforcement Learning (RL) systems has led to remarkable progress in solving complex tasks but also introduced challenges like primacy bias and dead neurons. Primacy bias skews learning towards early experiences, while dead neurons diminish the network's capacity to acquire new knowledge. Traditional reset mechanisms aimed at addressing these issues often involve maintaining large replay buffers to train new networks or selectively resetting subsets of neurons. However, these approaches either incur prohibitive computational costs or reset network parameters without ensuring stability through recovery mechanisms, ultimately impairing learning efficiency. In this work, we introduce the novel concept of neuron regeneration, which combines reset mechanisms with knowledge recovery techniques. We also propose a new framework called Sustainable Backup Propagation(SBP) that effectively maintains plasticity in neural networks through this neuron regeneration process. The SBP framework achieves whole network neuron regeneration through two key procedures: cycle reset and inner distillation. Cycle reset involves a scheduled renewal of neurons, while inner distillation functions as a knowledge recovery mechanism at the neuron level. To validate our framework, we integrate SBP with Proximal Policy Optimization (PPO) and propose a novel distillation function for inner distillation. This integration results in Plastic PPO (P3O), a new algorithm that enables efficient cyclic regeneration of all neurons in the actor network. Extensive experiments demonstrate the approach effectively maintains policy plasticity and improves sample efficiency in reinforcement learning.

Reinforcement learning (RL) has demonstrated compelling performance in robotic tasks, but its success often hinges on the design of complex, ad hoc reward functions. Researchers have explored how Large Language Models (LLMs) could enable non-expert users to specify reward functions more easily. However, LLMs struggle to balance the importance of different features, generalize poorly to out-of-distribution robotic tasks, and cannot represent the problem properly with only text-based descriptions. To address these challenges, we propose ELEMENTAL (intEractive LEarning froM dEmoNstraTion And Language), a novel framework that combines natural language guidance with visual user demonstrations to align robot behavior with user intentions better. By incorporating visual inputs, ELEMENTAL overcomes the limitations of text-only task specifications, while leveraging inverse reinforcement learning (IRL) to balance feature weights and match the demonstrated behaviors optimally. ELEMENTAL also introduces an iterative feedback-loop through self-reflection to improve feature, reward, and policy learning. Our experiment results demonstrate that ELEMENTAL outperforms prior work by 42.3% on task success, and achieves 41.3% better generalization in out-of-distribution tasks, highlighting its robustness in LfD.

In-Context Reinforcement Learning (ICRL) represents a promising paradigm for developing generalist agents that learn at inference time through trial-and-error interactions, analogous to how large language models adapt contextually, but with a focus on reward maximization. However, the scalability of ICRL beyond toy tasks and single-domain settings remains an open challenge. In this work, we present the first steps toward scaling ICRL by introducing a fixed, cross-domain model capable of learning behaviors through in-context reinforcement learning. Our results demonstrate that Algorithm Distillation, a framework designed to facilitate ICRL, offers a compelling and competitive alternative to expert distillation to construct versatile action models. These findings highlight the potential of ICRL as a scalable approach for generalist decision-making systems.

In this work, we introduce the Time-Aware World Model (TAWM), a model-based approach that explicitly incorporates temporal dynamics. By conditioning on the time-step size, $\Delta t$, and training over a diverse range of $\Delta t$ values – rather than sampling at a fixed time-step – TAWM learns both high- and low-frequency task dynamics across diverse control problems. Grounded in the information-theoretic insight that the optimal sampling rate depends on a system’s underlying dynamics, this time-aware formulation improves both performance and data efficiency. Empirical evaluations show that TAWM consistently outperforms conventional models across varying observation rates in a variety of control tasks, using the same number of training samples and iterations. Our code can be found online at: github.com/anh-nn01/Time-Aware-World-Model.

In recent years, contrastive learning has achieved state-of-the-art performance in the territory of self-supervised representation learning. Many previous works have attempted to provide the theoretical understanding underlying the success of contrastive learning. Almost all of them rely on a default assumption, i.e., the label consistency assumption, which may not hold in practice (the probability of failure is called labeling error) due to the strength and randomness of common augmentation strategies, such as random resized crop (RRC). This paper investigates the theoretical impact of labeling error on the downstream classification performance of contrastive learning. We first reveal several significant negative impacts of labeling error on downstream classification risk. To mitigate these impacts, data dimensionality reduction method (e.g., singular value decomposition, SVD) is applied on original data to reduce false positive samples, and establish both theoretical and empirical evaluations. Moreover, it is also found that SVD acts as a double-edged sword, which may lead to the deterioration of downstream classification accuracy due to the reduced connectivity of the augmentation graph. Based on the above observations, we give the augmentation suggestion that we should use some moderate embedding dimension (such as $512, 1024$ in our experiments), data inflation, weak augmentation, and SVD to ensure large graph connectivity and small labeling error to improve model performance.

Future advanced AI systems may learn sophisticated strategies through reinforcement learning (RL) that humans cannot understand well enough to safely evaluate. We propose a training method which avoids agents learning undesired multi-step plans that receive high reward (multi-step "reward hacks") even if humans are not able to detect that the behavior is undesired. The method, Myopic Optimization with Non-myopic Approval (MONA), works by combining short-sighted optimization with far-sighted reward. We demonstrate that MONA can prevent multi-step reward hacking that ordinary RL causes, even without being able to detect the reward hacking and without any extra information that ordinary RL does not get access to. We study MONA empirically in three settings which model different misalignment failure modes including 2-step environments with LLMs representing delegated oversight and encoded reasoning and longer-horizon gridworld environments representing sensor tampering.

The idea of value-aware model learning, that models should produce accurate value estimates, has gained prominence in model-based reinforcement learning. The MuZero loss, which penalizes a model's value function prediction compared to the ground-truth value function, has been utilized in several prominent empirical works in the literature. However, theoretical investigation into its strengths and weaknesses is limited. In this paper, we analyze the family of value-aware model learning losses, which includes the popular MuZero loss. We show that these losses, as normally used, are uncalibrated surrogate losses, which means that they do not always recover the correct model and value function. Building on this insight, we propose corrections to solve this issue. Furthermore, we investigate the interplay between the loss calibration, latent model architectures, and auxiliary losses that are commonly employed when training MuZero-style agents. We show that while deterministic models can be sufficient to predict accurate values, learning calibrated stochastic models is still advantageous.

Hierarchical reinforcement learning (HRL) improves the efficiency of long-horizon reinforcement-learning tasks with sparse rewards by decomposing the task into a hierarchy of subgoals. The main challenge of HRL is efficient discovery of the hierarchical structure among subgoals and utilizing this structure to achieve the final goal. We address this challenge by modeling the subgoal structure as a causal graph and propose a causal discovery algorithm to learn it. Additionally, rather than intervening on the subgoals at random during exploration, we harness the discovered causal model to prioritize subgoal interventions based on their importance in attaining the final goal. These targeted interventions result in a significantly more efficient policy in terms of the training cost. Unlike previous work on causal HRL, which lacked theoretical analysis, we provide a formal analysis of the problem. Specifically, for tree structures and, for a variant of Erdős-Rényi random graphs, our approach results in remarkable improvements. Our experimental results on HRL tasks also illustrate that our proposed framework outperforms existing work in terms of training cost.

Online safe reinforcement learning (RL) plays a key role in dynamic environments, with applications in autonomous driving, robotics, and cybersecurity. The objective is to learn optimal policies that maximize rewards while satisfying safety constraints modeled by constrained Markov decision processes (CMDPs). Existing methods achieve sublinear regret under stochastic constraints but often fail in adversarial settings, where constraints are unknown, time-varying, and potentially adversarially designed. In this paper, we propose the Optimistic Mirror Descent Primal-Dual (OMDPD) algorithm, the first to address online CMDPs with anytime adversarial constraints. OMDPD achieves optimal regret $\tilde{\mathcal{O}}(\sqrt{K})$ and strong constraint violation $\tilde{\mathcal{O}}(\sqrt{K})$ without relying on Slater’s condition or the existence of a strictly known safe policy. We further show that access to accurate estimates of rewards and transitions can further improve these bounds. Our results offer practical guarantees for safe decision-making in adversarial environments.

Standard multi-agent reinforcement learning (MARL) algorithms are vulnerable to sim-to-real gaps. To address this, distributionally robust Markov games (RMGs) have been proposed to enhance robustness in MARL by optimizing the worst-case performance when game dynamics shift within a prescribed uncertainty set. RMGs remains under-explored, from reasonable problem formulation to the development of sample-efficient algorithms. Two notorious and open challenges are the formulation of the uncertainty set and whether the corresponding RMGs can overcome the curse of multiagency, where the sample complexity scales exponentially with the number of agents. In this work, we propose a natural class of RMGs inspired by behavioral economics, where each agent's uncertainty set is shaped by both the environment and the integrated behavior of other agents. We first establish the well-posedness of this class of RMGs by proving the existence of game-theoretic solutions such as robust Nash equilibria and coarse correlated equilibria (CCE). Assuming access to a generative model, we then introduce a sample-efficient algorithm for learning the CCE whose sample complexity scales polynomially with all relevant parameters. To the best of our knowledge, this is the first algorithm to break the curse of multiagency for RMGs, regardless of the uncertainty set formulation.

We consider the problem of *exact* recovery of a $k$-sparse binary vector from generalized linear measurements (such as *logistic regression*). We analyze the *linear estimation* algorithm (Plan, Vershynin, Yudovina, 2017), and also show information theoretic lower bounds on the number of required measurements. As a consequence of our results, for noisy one bit quantized linear measurements ($\mathsf{1bCS}$), we obtain a sample complexity of $O((k+\sigma^2)\log{n})$, where $\sigma^2$ is the noise variance. This is shown to be optimal due to the information theoretic lower bound. We also obtain tight sample complexity characterization for logistic regression. Since $\mathsf{1bCS}$ is a strictly harder problem than noisy linear measurements ($\mathsf{SparseLinearReg}$) because of added quantization, the same sample complexity is achievable for $\mathsf{SparseLinearReg}$. While this sample complexity can be obtained via the popular lasso algorithm, linear estimation is computationally more efficient. Our lower bound holds for any set of measurements for $\mathsf{SparseLinearReg}$ (similar bound was known for Gaussian measurement matrices) and is closely matched by the maximum-likelihood upper bound. For $\mathsf{SparseLinearReg}$, it was conjectured in Gamarnik and Zadik, 2017 that there is a statistical-computational gap and the number of measurements should be at least $(2k+\sigma^2)\log{n}$ for efficient algorithms to exist. It is worth noting that our results imply that there is no such statistical-computational gap for $\mathsf{1bCS}$ and logistic regression.