Action-Dependent Optimality-Preserving Reward Shaping
Grant C. Forbes, Jianxun Wang, Leonardo Villalobos-Arias, Arnav Jhala, David L. Roberts
2025年份
摘要
Recent RL research has utilized reward shapingparticularly complex shaping rewards such as intrinsic motivation (IM)-to encourage agent exploration in sparse-reward environments. While
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了每一个公式，引用它的顶会 Paper 也一样。你提问，回答直接引用原文。
它引用的顶会 Paper4
- Never Give Up: Learning Directed Exploration StrategiesAdrià Puigdomènech Badia, Pablo Sprechmann, Alex Vitvitskyi, Zhaohan Daniel Guo 等ICLR 2020 · 被引用 349 次
- Redeeming intrinsic rewards via constrained optimizationEric Chen, Zhang-Wei Hong, Joni Pajarinen, Pulkit AgrawalNeurIPS 2022 · 被引用 48 次
- Automatic Intrinsic Reward Shaping for Exploration in Deep Reinforcement LearningMingqi Yuan, Bo Li, Xin Jin, Wenjun ZengICML 2023 · 被引用 17 次
- On Stationary Point Convergence of PPO-ClipRuinan Jin, Shuai Li, Baoxiang WangICLR 2024 · 被引用 14 次
相关 Paper
- Exploration-Guided Reward Shaping for Reinforcement Learning under Sparse RewardsRati Devidze, Parameswaran Kamalaruban, Adish SinglaNeurIPS 2022 · 被引用 122 次
- Mutual Information State Intrinsic ControlRui Zhao, Yang Gao, Pieter Abbeel, Volker Tresp 等ICLR 2021 · 被引用 25 次
- Sequential Generative Exploration Model for Partially Observable Reinforcement LearningHaiyan Yin, Jianda Chen, Sinno Jialin Pan, Sebastian TschiatschekAAAI 2021 · 被引用 7 次
- Intrinsic Motivation for Encouraging Synergistic BehaviorRohan Chitnis, Shubham Tulsiani, Saurabh Gupta, Abhinav GuptaICLR 2020 · 被引用 28 次
- RIDE: Rewarding Impact-Driven Exploration for Procedurally-Generated EnvironmentsRoberta Raileanu, Tim RocktäschelICLR 2020 · 被引用 198 次
