Research map / Reinforcement learning / MDPs and planning
Exploiting Multiple Abstractions in Episodic RL via Reward Shaping
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 abstract-MDP value estimates as non-return-invariant reward shaping | Q-learning · 8-rooms | episode length | Helped |
| Use abstract-MDP value estimates as reward shaping | Dueling DQN · Continuous interaction task | episode length | Helped |
| Use abstract-MDP value estimates as reward shaping | Q-learning · 4-rooms | episode length | Mixed |
| Set terminal-state shaping potentials to zero (ablation) | Q-learning with abstract-value reward shaping · 8-rooms | episode length | Hurt |
| Add an impossible abstract transition from the pink region to the goal (ablation) | Q-learning with abstract-value reward shaping · 8-rooms | episode length | Hurt |
| Also add an impossible abstract transition from the blue region to the pink region (ablation) | Q-learning with abstract-value reward shaping · 8-rooms | episode length | Hurt |
| Use a modeled abstraction containing errors rather than uninformed exploration (ablation) | Q-learning with abstract-value reward shaping · 8-rooms | episode length | Helped |
Closest papers by meaning
- Learning to Utilize Shaping Rewards: A New Approach of Reward Shaping (NeurIPS 2020 · similarity 0.91)
- Temporal-Logic-Based Reward Shaping for Continuing Reinforcement Learning Tasks (AAAI 2021 · similarity 0.90)
- Autonomous Option Invention for Continual Hierarchical Reinforcement Learning and Planning (AAAI 2025 · similarity 0.90)
- Context-Sensitive Abstractions for Reinforcement Learning with Parameterized Actions (AAAI 2026 · similarity 0.90)
- BAMDP Shaping: a Unified Framework for Intrinsic Motivation and Reward Shaping (ICLR 2025 · similarity 0.89)
- Reconciling Spatial and Temporal Abstractions for Goal Representation (ICLR 2024 · similarity 0.88)
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.