Neural Information Processing Systems http://nips.cc/

To improve the sample efficiency of policy-gradient based reinforcement learning algorithms, we propose implicit distributional actor-critic (IDAC) that consists of a distributional critic, built on two deep generator networks (DGNs), and a semi-implicit actor (SIA), powered by a flexible policy distribution. We adopt a distributional perspective on the discounted cumulative return and model it with a state-action-dependent implicit distribution, which is approximated by the DGNs that take state-action pairs and random noises as their input. Moreover, we use the SIA to provide a semi-implicit policy distribution, which mixes the policy parameters with a reparameterizable distribution that is not constrained by an analytic density function. In this way, the policy's marginal distribution is implicit, providing the potential to model complex properties such as covariance structure and skewness, but its parameter and entropy can still be estimated. We incorporate these features with an off-policy algorithm framework to solve problems with continuous action space and compare IDAC with state-of-the-art algorithms on representative OpenAI Gym environments. We observe that IDAC outperforms these baselines in most tasks.

Large language model post-training relies on reinforcement learning to improve model capability and alignment quality. However, the off-policy training paradigm introduces distribution shift, which often pushes the policy beyond the trust region, leading to training instabilities manifested as fluctuations in policy entropy and unstable gradients. Although PPO-Clip mitigates this issue through importance clipping, it still overlooks the global distributional shift of actions. To address these challenges, we propose using the entropy ratio between the current and previous policies as a new global metric that effectively quantifies the relative change in policy exploration throughout updates. Building on this metric, we introduce an \textbf{Entropy Ratio Clipping} (ERC) mechanism that imposes bidirectional constraints on the entropy ratio. This stabilizes policy updates at the global distribution level and compensates for the inability of PPO-clip to regulate probability shifts of un-sampled actions. We integrate ERC into both DAPO and GPPO reinforcement learning algorithms. Experiments across multiple benchmarks show that ERC consistently improves performance.

In reinforcement learning (RL), it is often advantageous to consider additional constraints on the action space to ensure safety or action relevance. Existing work on such action-constrained RL faces challenges regarding effective policy updates, computational efficiency, and predictable runtime. Recent work proposes to use truncated normal distributions for stochastic policy gradient methods. However, the computation of key characteristics, such as the entropy, log-probability, and their gradients, becomes intractable under complex constraints. Hence, prior work approximates these using the non-truncated distributions, which severely degrades performance. We argue that accurate estimation of these characteristics is crucial in the action-constrained RL setting, and propose efficient numerical approximations for them. We also provide an efficient sampling strategy for truncated policy distributions and validate our approach on three benchmark environments, which demonstrate significant performance improvements when using accurate estimations.

Neural Information Processing Systems http://nips.cc/

End-to-end autonomous driving has substantially progressed by directly predicting future trajectories from raw perception inputs, which bypasses traditional modular pipelines. However, mainstream methods trained via imitation learning suffer from critical safety limitations, as they fail to distinguish between trajectories that appear human-like but are potentially unsafe. Some recent approaches attempt to address this by regressing multiple rule-driven scores but decoupling supervision from policy optimization, resulting in suboptimal performance. To tackle these challenges, we propose DriveDPO, a Safety Direct Preference Optimization Policy Learning framework. First, we distill a unified policy distribution from human imitation similarity and rule-based safety scores for direct policy optimization. Further, we introduce an iterative Direct Preference Optimization stage formulated as trajectory-level preference alignment. Extensive experiments on the NAVSIM benchmark demonstrate that DriveDPO achieves a new state-of-the-art PDMS of 90.0. Furthermore, qualitative results across diverse challenging scenarios highlight DriveDPO's ability to produce safer and more reliable driving behaviors.

Objective: This study aims to uncover the opaque decision-making process of an artificial intelligence (AI) agent for automatic treatment planning. Approach: We examined a previously developed AI agent based on the Actor-Critic with Experience Replay (ACER) network, which automatically tunes treatment planning parameters (TPPs) for inverse planning in prostate cancer intensity modulated radiotherapy. We selected multiple checkpoint ACER agents from different stages of training and applied an explainable AI (EXAI) method to analyze the attribution from dose-volume histogram (DVH) inputs to TPP-tuning decisions. We then assessed each agent's planning efficacy and efficiency and evaluated their policy and final TPP tuning spaces. Combining these analyses, we systematically examined how ACER agents generated high-quality treatment plans in response to different DVH inputs. Results: Attribution analysis revealed that ACER agents progressively learned to identify dose-violation regions from DVH inputs and promote appropriate TPP-tuning actions to mitigate them. Organ-wise similarities between DVH attributions and dose-violation reductions ranged from 0.25 to 0.5 across tested agents. Agents with stronger attribution-violation similarity required fewer tuning steps (~12-13 vs. 22), exhibited a more concentrated TPP-tuning space with lower entropy (~0.3 vs. 0.6), converged on adjusting only a few TPPs, and showed smaller discrepancies between practical and theoretical tuning steps. Putting together, these findings indicate that high-performing ACER agents can effectively identify dose violations from DVH inputs and employ a global tuning strategy to achieve high-quality treatment planning, much like skilled human planners. Significance: Better interpretability of the agent's decision-making process may enhance clinician trust and inspire new strategies for automatic treatment planning.

Offline cooperative multi-agent reinforcement learning (MARL) faces unique challenges due to distributional shifts, particularly stemming from the high dimensionality of joint action spaces and the presence of out-of-distribution joint action selections. In this work, we highlight that a fundamental challenge in offline MARL arises from the multi-equilibrium nature of cooperative tasks, which induces a highly multimodal joint behavior policy space coupled with heterogeneous-quality behavior data. This makes it difficult for individual policy regularization to align with a consistent coordination pattern, leading to the policy distribution shift problems. To tackle this challenge, we design a sequential score function decomposition method that distills per-agent regularization signals from the joint behavior policy, which induces coordinated modality selection under decentralized execution constraints. Then we leverage a flexible diffusion-based generative model to learn these score functions from multimodal offline data, and integrate them into joint-action critics to guide policy updates toward high-reward, in-distribution regions under a shared team reward. Our approach achieves state-of-the-art performance across multiple particle environments and Multi-agent MuJoCo benchmarks consistently. To the best of our knowledge, this is the first work to explicitly address the distributional gap between offline and online MARL, paving the way for more generalizable offline policy-based MARL methods.