PulseAugur
EN
LIVE 11:11:10

New research enhances diffusion models for robust RL and safe planning

By PulseAugur Editorial · [7 sources] ·

Researchers are developing new methods to improve the robustness and safety of diffusion models in reinforcement learning and planning tasks. One approach, Robust Regularized Policy Iteration (RRPI), addresses transition uncertainty by optimizing against worst-case dynamics and has shown strong performance on D4RL benchmarks. Another set of papers introduces techniques like Kolmogorov Regression and DiRecT to enhance diffusion policies by improving trajectory regularity, enabling deterministic failure detection, and enforcing safety constraints during inference without over-constraining the sampling process. These advancements aim to make diffusion models more reliable for complex, long-horizon tasks and safety-critical applications. AI

IMPACT These advancements aim to improve the reliability and safety of AI systems in complex decision-making tasks.

RANK_REASON Cluster consists of multiple academic papers on novel algorithms for reinforcement learning and diffusion models.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 7 sources. How we write summaries →

New research enhances diffusion models for robust RL and safe planning

COVERAGE [7]

  1. arXiv cs.AI TIER_1 English(EN) · Zhengyi Guo, Wenpin Tang, Renyuan Xu ·

    Conditional Diffusion Guidance under Hard Constraint: A Stochastic Analysis Approach

    arXiv:2602.05533v3 Announce Type: replace Abstract: We study conditional generation in diffusion models under hard constraints, where generated samples must satisfy prescribed events with probability one. Such constraints arise naturally in safety-critical applications and in rar…

  2. arXiv cs.AI TIER_1 English(EN) · Hongqiang Lin, Zhenghui Fu, Weihao Tang, Pengfei Wang, Yiding Sun, Qixian Huang, Dongxu Zhang ·

    Robust Regularized Policy Iteration under Transition Uncertainty

    arXiv:2603.09344v3 Announce Type: replace Abstract: Offline reinforcement learning (RL) enables data-efficient and safe policy learning without online exploration, but its performance often degrades under distribution shift. The learned policy may visit out-of-distribution state-…

  3. arXiv cs.AI TIER_1 English(EN) · Lekan Molu ·

    Kolmogorov Regression for Robust Diffusion Policies

    arXiv:2606.18186v1 Announce Type: cross Abstract: Finite-dimensional (FD) diffusion policies exhibit temporal drift owing to discretization artifacts that degrade long-horizon performance (when deployed on physical systems). We introduce a backward Kolmogorov equation that lifts …

  4. arXiv cs.AI TIER_1 English(EN) · Lekan Molu ·

    Kolmogorov Regression for Robust Diffusion Policies

    Finite-dimensional (FD) diffusion policies exhibit temporal drift owing to discretization artifacts that degrade long-horizon performance (when deployed on physical systems). We introduce a backward Kolmogorov equation that lifts diffusion policies to a Cameron-Martin space -- a …

  5. arXiv cs.LG TIER_1 English(EN) · Paolo Giaretta, Zeyang Li, Navid Azizan ·

    DiRecT: Safe Diffusion-Based Planning via Receding-Horizon Denoising

    arXiv:2606.15359v1 Announce Type: new Abstract: Diffusion models have emerged as powerful tools for planning and control by learning multimodal distributions over actions and trajectories. Yet reliable inference-time safety enforcement remains a key barrier to their deployment in…

  6. arXiv cs.AI TIER_1 English(EN) · Abhinav Agarwal, Adam Wei, Taylan Kargin, Michael Zeng, Cole Becker, Arif Kerem Dayi, Pablo Parrilo, Asuman Ozdaglar, Russ Tedrake ·

    Training and Evaluating Diffusion Policies with Long Context Lengths

    arXiv:2606.16447v1 Announce Type: cross Abstract: Imitation learning has enabled highly-dexterous robotic manipulation from RGB observations. Policies trained with these methods, however, typically condition robot actions on only a short history of observations. These policies ca…

  7. arXiv cs.LG TIER_1 English(EN) · Haozhe Jiang, Haiwen Feng, Pieter Abbeel, Jiantao Jiao, Angjoo Kanazawa, Nika Haghtalab ·

    Diffusion Policy Optimization without Drifting Apart

    arXiv:2606.13795v1 Announce Type: new Abstract: RL post-training has become increasingly pivotal for improving diffusion policies, but existing diffusion policy-gradient methods are often unstable and cannot achieve reliable policy improvement. We identify the cause as the double…

RELATED ENTITIES

RELATED TOPICS