Research map / Reinforcement learning / Visual and representation RL
Learning Long-Term Reward Redistribution via Randomized Return Decomposition
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 |
|---|---|---|---|
| Use randomized return decomposition with subsequence length K=64 instead of the compared baselines | MuJoCo locomotion tasks with episodic rewards · MuJoCo locomotion tasks | sample efficiency and policy quality | Helped |
| Use randomized return decomposition instead of the compared baselines | Atari DQN with episodic rewards · Atari games | performance | Mixed |
| Add the Monte-Carlo variance penalty by optimizing L_Rand-RD rather than the unbiased L_RD objective (ablation) | RRD · MuJoCo locomotion tasks | sample efficiency and policy quality | Helped |
| Increase subsequence length K (ablation) | RRD · MuJoCo locomotion tasks | policy performance | Mixed |
| Use the cited IRCR hyperparameters rather than the alternative SAC configuration (ablation) | IRCR · MuJoCo locomotion tasks | training stability | Helped |
Closest papers by meaning
- Episodic Return Decomposition by Difference of Implicitly Assigned Sub-trajectory Reward (AAAI 2024 · similarity 0.92)
- The Distributional Reward Critic Framework for Reinforcement Learning Under Perturbed Rewards (AAAI 2025 · similarity 0.87)
- DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning (NeurIPS 2025 · similarity 0.86)
- In-Dataset Trajectory Return Regularization for Offline Preference-based Reinforcement Learning (AAAI 2025 · similarity 0.86)
- Reinforcement Learning with Action Chunking (NeurIPS 2025 · similarity 0.86)
- Episodic Novelty Through Temporal Distance (ICLR 2025 · similarity 0.86)
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.