Research map / Reinforcement learning / Imitation learning
Efficient Exploration of Reward Functions in Inverse Reinforcement Learning via Bayesian Optimization
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 the ρ-RBF kernel instead of standard RBF or Matérn kernels (ablation) | BO-IRL · Gridworld | iterations to achieve expert's ESOR: 30.0±34.4 -> 16.0±15.6; 22.2±12.2 -> 16.0±15.6 | Helped |
| Use the ρ-RBF kernel instead of RBF (ablation) | BO-IRL · Börlange road network | iterations to achieve expert's ESOR: 9.5±6.3 -> 2.0±1.1 | Helped |
| Use the ρ-RBF kernel instead of Matérn (ablation) | BO-IRL · Börlange road network | iterations to achieve expert's ESOR: 5.6±3.8 -> 2.0±1.1 | Helped |
| Use the ρ-RBF kernel instead of RBF (ablation) | BO-IRL · Point Mass Maze | success rate: 20% -> 80% | Helped |
| Use the ρ-RBF kernel instead of Matérn (ablation) | BO-IRL · Point Mass Maze | success rate: 20% -> 80% | Helped |
| Use the ρ-RBF kernel instead of Matérn (ablation) | BO-IRL · Point Mass Maze | iterations to achieve expert's ESOR: 56±29 -> 51.4±23.1 | Hurt |
| Use BIRL instead of BO-IRL with ρ-RBF | Gridworld · Gridworld | success rate: 70% -> 80% | Helped |
| Use ρ-projection to account for policy invariance in the reward-space kernel (ablation) | BO-IRL · Gridworld, Börlange road network, and Point Mass Maze | iterations to achieve expert's ESOR: 30.0±34.4 -> 16.0±15.6; 9.5±6.3 -> 2.0±1.1; 28.0±4 -> 51.4±23.1 | Helped |
Closest papers by meaning
- Recent Advancements in Inverse Reinforcement Learning (AAAI 2024 · similarity 0.92)
- A Geometric Perspective on Reward Function Updates in Inverse Reinforcement Learning (NeurIPS 2026 · similarity 0.91)
- Is Inverse Reinforcement Learning Harder than Standard Reinforcement Learning? A Theoretical Perspective (ICML 2024 · similarity 0.91)
- How does Inverse RL Scale to Large State Spaces? A Provably Efficient Approach (NeurIPS 2024 · similarity 0.91)
- A Unified Linear Programming Framework for Offline Reward Learning from Human Demonstrations and Feedback (ICML 2024 · similarity 0.90)
- Offline Inverse RL: New Solution Concepts and Provably Efficient Algorithms (ICML 2024 · similarity 0.90)
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.