We introduce an offline multi-agent reinforcement learning ( offline MARL) framework that utilizes previously collected data without additional online data collection. Our method reformulates offline MARL as a sequence modeling problem and thus builds on top of the simplicity and scalability of the Transformer architecture. In the fashion of centralized training and decentralized execution, we propose to first train a teacher policy as if the MARL dataset is generated by a single agent. After the teacher policy has identified and recombined the good behavior in the dataset, we create separate student policies and distill not only the teacher policy's features but also its structural relations among different agents' features to student policies. Despite its simplicity, the proposed method outperforms state-of-the-art model-free offline MARL baselines while being more robust to demonstration's quality on several environments.

We introduce the forward-backward (FB) representation of the dynamics of a reward-free Markov decision process. It provides explicit near-optimal policies for any reward specified a posteriori. During an unsupervised phase, we use reward-free interactions with the environment to learn two representations via off-the-shelf deep learning methods and temporal difference (TD) learning. In the test phase, a reward representation is estimated either from reward observations or an explicit reward description (e.g., a target state). The optimal policy for thatreward is directly obtained from these representations, with no planning.

Strong inductive biases give humans the ability to quickly learn to perform a variety of tasks. Although meta-learning is a method to endow neural networks with useful inductive biases, agents trained by meta-learning may sometimes acquire very different strategies from humans. We show that co-training these agents on predicting representations from natural language task descriptions and programs induced to generate such tasks guides them toward more human-like inductive biases. Human-generated language descriptions and program induction models that add new learned primitives both contain abstract concepts that can compress description length. Co-training on these representations result in more human-like behavior in downstream meta-reinforcement learning agents than less abstract controls (synthetic language descriptions, program induction without learned primitives), suggesting that the abstraction supported by these representations is key.

We provide improved gap-dependent regret bounds for reinforcement learning in finite episodic Markov decision processes. Compared to prior work, our bounds depend on alternative definitions of gaps. These definitions are based on the insight that, in order to achieve a favorable regret, an algorithm does not need to learn how to behave optimally in states that are not reached by an optimal policy. We prove tighter upper regret bounds for optimistic algorithms and accompany them with new information-theoretic lower bounds for a large class of MDPs. Our results show that optimistic algorithms can not achieve the information-theoretic lower bounds even in deterministic MDPs unless there is a unique optimal policy.

Emphatic algorithms have shown great promise in stabilizing and improving reinforcement learning by selectively emphasizing the update rule. Although the emphasis fundamentally depends on an interest function which defines the intrinsic importance of each state, most approaches simply adopt a uniform interest over all states (except where a hand-designed interest is possible based on domain knowledge). In this paper, we investigate adaptive methods that allow the interest function to dynamically vary over states and iterations. In particular, we leverage meta-gradients to automatically discover online an interest function that would accelerate the agent's learning process. Empirical evaluations on a wide range of environments show that adapting the interest is key to provide significant gains. Qualitative analysis indicates that the learned interest function emphasizes states of particular importance, such as bottlenecks, which can be especially useful in a transfer learning setting.

Multi-agent reinforcement learning (MARL) research is faced with a trade-off: it either uses complex environments requiring large compute resources, which makes it inaccessible to researchers with limited resources, or relies on simpler dynamics for faster execution, which makes the transferability of the results to more realistic tasks challenging. Motivated by these challenges, we present Gigastep, a fully vectorizable, MARL environment implemented in JAX, capable of executing up to one billion environment steps per second on consumer-grade hardware. Its design allows for comprehensive MARL experimentation, including a complex, high-dimensional space defined by 3D dynamics, stochasticity, and partial observations. Gigastep supports both collaborative and adversarial tasks, continuous and discrete action spaces, and provides RGB image and feature vector observations, allowing the evaluation of a wide range of MARL algorithms.

We study robust reinforcement learning (RL) with the goal of determining a well-performing policy that is robust against model mismatch between the training simulator and the testing environment. Previous policy-based robust RL algorithms mainly focus on the tabular setting under uncertainty sets that facilitate robust policy evaluation, but are no longer tractable when the number of states scales up. To this end, we propose two novel uncertainty set formulations, one based on double sampling and the other on an integral probability metric. Both make large-scale robust RL tractable even when one only has access to a simulator. We propose a robust natural actor-critic (RNAC) approach that incorporates the new uncertainty sets and employs function approximation. We provide finite-time convergence guarantees for the proposed RNAC algorithm to the optimal robust policy within the function approximation error. Finally, we demonstrate the robust performance of the policy learned by our proposed RNAC approach in multiple MuJoCo environments and a real-world TurtleBot navigation task.

In this post, I'll introduce a reinforcement learning (RL) algorithm based on an "alternative" paradigm: divide and conquer We can do Reinforcement Learning (RL) based on divide and conquer, instead of temporal difference (TD) learning. There are two classes of algorithms in RL: on-policy RL and off-policy RL. On-policy RL means we can use fresh data collected by the current policy. In other words, we have to throw away old data each time we update the policy. Algorithms like PPO and GRPO (and policy gradient methods in general) belong to this category.

We present a novel theoretical analysis of Federated SARSA (FedSARSA) with linear function approximation and local training. We establish convergence guarantees for FedSARSA in the presence of heterogeneity, both in local transitions and rewards, providing the first sample and communication complexity bounds in this setting. At the core of our analysis is a new, exact multi-step error expansion for single-agent SARSA, which is of independent interest. Our analysis precisely quantifies the impact of heterogeneity, demonstrating the convergence of FedSARSA with multiple local updates. Crucially, we show that FedSARSA achieves linear speed-up with respect to the number of agents, up to higher-order terms due to Markovian sampling. Numerical experiments support our theoretical findings.

We introduce Stackelberg Learning from Human Feedback (SLHF), a new framework for preference optimization. SLHF frames the alignment problem as a sequential-move game between two policies: a Leader, which commits to an action, and a Follower, which responds conditionally on the Leader's action. This approach decomposes preference optimization into a refinement problem for the Follower and an optimization problem against an adversary for the Leader. Unlike Reinforcement Learning from Human Feedback (RLHF), which assigns scalar rewards to actions, or Nash Learning from Human Feedback (NLHF), which seeks a simultaneous-move equilibrium, SLHF leverages the asymmetry of sequential play to capture richer preference structures. The sequential design of SLHF naturally enables inference-time refinement, as the Follower learns to improve the Leader's actions, and these refinements can be leveraged through iterative sampling. We compare the solution concepts of SLHF, RLHF, and NLHF, and lay out key advantages in consistency, data sensitivity, and robustness to intransitive preferences. Experiments on large language models demonstrate that SLHF achieves strong alignment across diverse preference datasets, scales from 0.5B to 8B parameters, and yields inference-time refinements that transfer across model families without further fine-tuning.