The paper proposes a method for learning task-relevant representations in deep reinforcement learning by maximizing rollout total correlation, which captures the correlation among all learned representations and actions across entire trajectories. It introduces two complementary lower bounds—one generative and one discriminative—along with chunk‑wise mini‑batching to improve this objective, and also proposes an intrinsic reward derived from the learned representation to enhance exploration. Experiments on challenging image‑based simulated control tasks demonstrate improved sample efficiency and robustness to white noise and natural video backgrounds compared to leading baselines.
By Bang You, Huaping Liu, Jan Peters, Oleg Arenz
arXiv:2606. 05888v1 Announce Type: new Abstract: Retry-based objectives such as pass@K and max@K optimize the best return obtained from multiple sampled trajectories, and recent work has shown that they can promote exploration without explicit exploration bonuses.
By Soichiro Nishimori, Paavo Parmas
The paper introduces Diffusion Skill Discovery (DSD), a method that employs a diffusion model to approximate the entropy gradient of policy-induced state distributions via score matching. This approach encourages the learning of a diverse repertoire of motor skills for high‑dimensional humanoid control, overcoming limitations of prior mutual‑information based methods that rely on indirect state entropy estimates. The discovered skills are effectively reused in hierarchical control and zero‑shot tasks, yielding more complex and agile behaviors than previous skill discovery techniques.
By Sun Woo Kim, Xue Bin Peng
arXiv:2510. 12363v4 Announce Type: replace-cross Abstract: The pretraining-finetuning paradigm has facilitated numerous transformative advancements in artificial intelligence research in recent years.
By Jiale Fan, Andrei Cramariuc, Tifanny Portela, Marco Hutter
arXiv:2510. 11103v3 Announce Type: replace-cross Abstract: Many robotic control tasks require policies to act on orientations, yet the geometry of SO(3) makes this nontrivial.
By Martin Schuck, Sherif Samy, Angela P. Schoellig
arXiv:2605. 11020v2 Announce Type: replace-cross Abstract: Inverse reinforcement learning (IRL) is typically formulated as maximizing entropy subject to matching the distribution of expert trajectories.
By Anish Diwan, Davide Tateo, Christopher E. Mower, Haitham Bou-Ammar, Jan Peters, Oleg Arenz
arXiv:2512. 07212v3 Announce Type: replace Abstract: Imitation learning with diffusion models has advanced robotic control by capturing the multi-modal action distributions.
By Zhaoyang Liu, Mokai Pan, Zhongyi Wang, Kaizhen Zhu, Haotao Lu, Haipeng Zhang, Jingya Wang, Ye Shi
arXiv:2606. 00083v1 Announce Type: cross Abstract: Reinforcement learning relies on accurate reward functions, which are often hand-crafted or even unavailable in real-world applications, such as robotics.
By Christian Gumbsch, Leonardo Barcellona, Lennard Sch\"unemann, Platon Karageorgis, Andrii Zadaianchuk, Zehao Wang, Sergey Zakharov, Fabien Despinoy, Rahaf Aljundi, Efstratios Gavves
arXiv:2601. 23075v2 Announce Type: replace Abstract: On-policy Reinforcement Learning (RL) remains a dominant paradigm for continuous control, yet standard implementations rely on Gaussian actors and relatively shallow MLP policies, often leading to brittle optimization when gradients are noisy, and policy updates must be conservative.
By Yuexin Bian, Jie Feng, Tao Wang, Yijiang Li, Sicun Gao, Yuanyuan Shi
arXiv:2603. 25464v2 Announce Type: replace-cross Abstract: Zero-shot reinforcement learning (RL) algorithms aim to learn a family of policies from a reward-free dataset, and recover optimal policies for any reward function directly at test time.
By Jiajun Hu, Nuria Armengol Urpi, Jin Cheng, Stelian Coros
arXiv:2607. 21302v1 Announce Type: new Abstract: Behavior prior reinforcement learning (BPRL) has emerged as a promising paradigm to improve sample efficiency in online reinforcement learning (RL) by leveraging policy priors derived from offline demonstrations.
By Gong Gao, Weidong Zhao, Xianhui Liu, Ning Jia
arXiv:2606. 32027v1 Announce Type: cross Abstract: Reward design remains a central bottleneck for autonomous robot policy improvement, especially in long-horizon manipulation tasks where sparse success labels provide too little signal and binary preferences collapse many competing notions of quality into one ambiguous signal.
By Marcel Torne, Anubha Mahajan, Abhijnya Bhat, Chelsea Finn