Research map / Reinforcement learning / Policy gradient and actor-critic
Instability of Meta-Learning Intrinsic Rewards for Policy Gradient Reinforcement Learning
Dilith Jayakody, Domenic Rosati, Janarthanan Rajendran
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Examines the instability of meta-learned intrinsic rewards in the LIRPG framework for policy gradient RL, based on the abstract alone.
Problem
Meta-learned intrinsic rewards (LIRPG and its successors) are used to shape policy learning when extrinsic rewards are sparse, delayed, or unavailable, but their training stability is in question.
Approach
Studies the instability of LIRPG-style meta-learned intrinsic rewards for policy gradient RL, the foundation for subsequent methods; the abstract does not state the specific fix or finding.
Experimental setup
Builds on: LIRPG
Closest papers by meaning
- Learning to Generalize from Sparse and Underspecified Rewards (ICML 2019 · similarity 0.88)
- What Can Learned Intrinsic Rewards Capture? (ICML 2020 · similarity 0.88)
- MetaCURE: Meta Reinforcement Learning with Empowerment-Driven Exploration (ICML 2021 · similarity 0.88)
- ProMP: Proximal Meta-Policy Search (ICLR 2019 · similarity 0.88)
- Enhanced Meta Reinforcement Learning via Demonstrations in Sparse Reward Environments (NeurIPS 2022 · similarity 0.88)
- MetaCARD: Meta-Reinforcement Learning with Task Uncertainty Feedback via Decoupled Context-Aware Reward and Dynamics Components (AAAI 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
