中文
相关论文

相关论文: Goal Exploration Augmentation via Pre-trained Skil…

200 篇论文

Exploration efficiency poses a significant challenge in goal-conditioned reinforcement learning (GCRL) tasks, particularly those with long horizons and sparse rewards. A primary limitation to exploration efficiency is the agent's inability…

机器学习 · 计算机科学 2024-04-22 Lisheng Wu , Ke Chen

In this work, we take a first step toward elucidating the mechanisms behind emergent exploration in unsupervised reinforcement learning. We study Single-Goal Contrastive Reinforcement Learning (SGCRL), a self-supervised algorithm capable of…

机器学习 · 计算机科学 2025-10-17 Mahsa Bastankhah , Grace Liu , Dilip Arumugam , Thomas L. Griffiths , Benjamin Eysenbach

Goal-conditioned hierarchical reinforcement learning (GCHRL) provides a promising approach to solving long-horizon tasks. Recently, its success has been extended to more general settings by concurrently learning hierarchical policies and…

机器学习 · 计算机科学 2022-03-08 Siyuan Li , Jin Zhang , Jianhao Wang , Yang Yu , Chongjie Zhang

Goal-Conditioned Reinforcement Learning (GCRL) provides a versatile framework for developing unified controllers capable of handling wide ranges of tasks, exploring environments, and adapting behaviors. However, its reliance on…

机器学习 · 计算机科学 2025-02-20 Charly Pecqueux-Guézénec , Stéphane Doncieux , Nicolas Perrin-Gilbert

Unsupervised goal-conditioned reinforcement learning (GCRL) is a promising paradigm for developing diverse robotic skills without external supervision. However, existing unsupervised GCRL methods often struggle to cover a wide range of…

机器学习 · 计算机科学 2024-12-10 Junik Bae , Kwanyoung Park , Youngwoon Lee

Goal-Conditioned Reinforcement Learning (GCRL) can enable agents to spontaneously set diverse goals to learn a set of skills. Despite the excellent works proposed in various fields, reaching distant goals in temporally extended tasks…

机器人学 · 计算机科学 2023-07-21 Zhifeng Qian , Mingyu You , Hongjun Zhou , Xuanhui Xu , Bin He

Offline goal-conditioned reinforcement learning (GCRL) often struggles with long-horizon tasks, where errors in value estimation accumulate and produce unreliable policies. It is typically assumed that effective long-term planning is…

机器学习 · 计算机科学 2026-05-26 Evgenii Opryshko , Junwei Quan , Claas Voelcker , Yilun Du , Igor Gilitschenski

One of the gnarliest challenges in reinforcement learning (RL) is exploration that scales to vast domains, where novelty-, or coverage-seeking behaviour falls short. Goal-directed, purposeful behaviours are able to overcome this, but rely…

机器学习 · 计算机科学 2023-02-10 Akhil Bagaria , Ray Jiang , Ramana Kumar , Tom Schaul

Hierarchical Reinforcement Learning (HRL) exploits temporally extended actions, or options, to make decisions from a higher-dimensional perspective to alleviate the sparse reward problem, one of the most challenging problems in…

机器学习 · 计算机科学 2019-05-15 Libo Xing

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…

机器学习 · 计算机科学 2025-07-08 Sadegh Khorasani , Saber Salehkaleybar , Negar Kiyavash , Matthias Grossglauser

Multi-goal reinforcement learning (RL) aims to qualify the agent to accomplish multi-goal tasks, which is of great importance in learning scalable robotic manipulation skills. However, reward engineering always requires strenuous efforts in…

机器人学 · 计算机科学 2021-09-27 Deyu Yang , Hanbo Zhang , Xuguang Lan , Jishiyu Ding

We propose a hierarchical entity-centric framework for offline Goal-Conditioned Reinforcement Learning (GCRL) that combines subgoal decomposition with factored structure to solve long-horizon tasks in domains with multiple entities.…

机器学习 · 计算机科学 2026-02-04 Dan Haramati , Carl Qi , Tal Daniel , Amy Zhang , Aviv Tamar , George Konidaris

Real-world decision-making tasks typically occur in complex and open environments, posing significant challenges to reinforcement learning (RL) agents' exploration efficiency and long-horizon planning capabilities. A promising approach is…

机器学习 · 计算机科学 2025-09-29 Yajie Qi , Wei Wei , Lin Li , Lijun Zhang , Zhidong Gao , Da Wang , Huizhong Song

Goal-conditioned reinforcement learning (GCRL), related to a set of complex RL problems, trains an agent to achieve different goals under particular scenarios. Compared to the standard RL solutions that learn a policy solely depending on…

人工智能 · 计算机科学 2022-09-05 Minghuan Liu , Menghui Zhu , Weinan Zhang

Recently, a simple yet effective algorithm -- goal-conditioned supervised-learning (GCSL) -- was proposed to tackle goal-conditioned reinforcement-learning. GCSL is based on the principle of hindsight learning: by observing states visited…

机器学习 · 计算机科学 2023-05-18 Tom Jurgenson , Aviv Tamar

Sparse-reward reinforcement learning (RL) can model a wide range of highly complex tasks. Solving sparse-reward tasks is RL's core premise, requiring efficient exploration coupled with long-horizon credit assignment, and overcoming these…

机器学习 · 计算机科学 2025-10-21 Leander Diaz-Bone , Marco Bagatella , Jonas Hübotter , Andreas Krause

Offline goal-conditioned reinforcement learning (GCRL) aims at solving goal-reaching tasks with sparse rewards from an offline dataset. While prior work has demonstrated various approaches for agents to learn near-optimal policies, these…

机器人学 · 计算机科学 2024-03-05 Chenyang Cao , Zichen Yan , Renhao Lu , Junbo Tan , Xueqian Wang

Offline goal-conditioned reinforcement learning (GCRL) is a promising approach for pretraining generalist policies on large datasets of reward-free trajectories, akin to the self-supervised objectives used to train foundation models for…

机器学习 · 计算机科学 2026-01-05 John L. Zhou , Jonathan C. Kao

Offline goal-conditioned reinforcement learning (GCRL) is challenging in long-horizon tasks, where distant state--goal pairs provide weak supervision and value estimates become vulnerable to accumulated bootstrapping errors. Hierarchical…

机器学习 · 计算机科学 2026-05-28 Kaiqiang Ke , Shenghong He , Chengdong Xu , Yuheng Luo , Xiangyuan Lan , Chao Yu

Designing rewards for Reinforcement Learning (RL) is challenging because it needs to convey the desired task, be efficient to optimize, and be easy to compute. The latter is particularly problematic when applying RL to robotics, where…

机器学习 · 计算机科学 2020-05-28 Yiming Ding , Carlos Florensa , Mariano Phielipp , Pieter Abbeel
‹ 上一页 1 2 3 10 下一页 ›