Research map / Reinforcement learning / Policy gradient and actor-critic
Meta-Gradient Reinforcement Learning with an Objective Discovered Online
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| learn the update target online with meta-gradients | FRODO actor-critic · Atari ALE, 57 games | median human-normalised score | Helped |
| learn a target from three-transition trajectories instead of using a fixed three-step lookahead | FRODO · Catch | average episode return | Helped |
| adapt the update target online to the changing reward | FRODO · 5-state Random Walk | mean squared prediction error | Helped |
| parameterise the update target rather than the loss (ablation) | FRODO · Atari ALE, 8-game subset | median human-normalised score | Helped |
| replace the target meta-network with a loss meta-network (ablation) | FRODO · Atari ALE, 8-game subset | median human-normalised score | Hurt |
| add a modest self-consistency loss with c=0.1 (ablation) | FRODO · Atari ALE, 8-game subset | median human-normalised score | Helped |
| remove the consistency loss (ablation) | FRODO · Atari ALE, 8-game subset | median human-normalised score | Hurt |
| increase consistency-loss weight to c=1 (ablation) | FRODO · Atari ALE, 8-game subset | median human-normalised score | Hurt |
Closest papers by meaning
- Bootstrapped Meta-Learning (ICLR 2022 · similarity 0.91)
- Discovering Temporally-Aware Reinforcement Learning Algorithms (ICLR 2024 · similarity 0.91)
- Can Learned Optimization Make Reinforcement Learning Less Difficult? (NeurIPS 2024 · similarity 0.90)
- AMAGO-2: Breaking the Multi-Task Barrier in Meta-Reinforcement Learning with Transformers (NeurIPS 2024 · similarity 0.89)
- Enhancing Online Reinforcement Learning with Meta-Learned Objective from Offline Data (AAAI 2025 · similarity 0.89)
- Meta-Reinforcement Learning with Universal Policy Adaptation: Provable Near-Optimality under All-task Optimum Comparator (NeurIPS 2024 · similarity 0.89)
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.