Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Repeated Inverse Reinforcement Learning
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 central-cut ellipsoid updates after each surprising task | Ellipsoid Algorithm for Repeated IRL · Linear bandit, adversarial tasks | mistakes: O(d^2 \log(d/\epsilon)) | Helped |
| Let the agent choose tasks and manipulate task rewards for binary search | Reward-identification algorithm · Agent-chosen tasks | demonstrations: O(\log(1/\epsilon)) | Helped |
| Use mini-batches of trajectory demonstrations rather than full policy demonstrations | Trajectory version of Ellipsoid Algorithm · MDPs, trajectory demonstrations | mistakes: \tilde O(\frac{d^2}{\epsilon^2} \log(\frac{d}{\delta \epsilon})) | Helped |
| Have nature choose tasks adversarially instead of allowing arbitrary agent-chosen tasks | Repeated IRL · Linear bandit, adversarial tasks | mistakes: \Omega(d \log(1/\epsilon)) | Hurt |
| Require positive feature diversity, measured by spread(X), in a fixed environment | Ellipsoid Algorithm with chosen task rewards · Fixed feature matrix X | identification error: \|c_{T} - \theta_\star\|_\infty \le \epsilon \sqrt{(K-1)/2} / \textrm{spread}(X) | Helped |
Closest papers by meaning
- Hybrid Inverse Reinforcement Learning (ICML 2024 · similarity 0.87)
- Recent Advancements in Inverse Reinforcement Learning (AAAI 2024 · similarity 0.87)
- Is Inverse Reinforcement Learning Harder than Standard Reinforcement Learning? A Theoretical Perspective (ICML 2024 · similarity 0.86)
- Imitation Learning from Observation with Automatic Discount Scheduling (ICLR 2024 · similarity 0.86)
- Rethinking Inverse Reinforcement Learning: from Data Alignment to Task Alignment (NeurIPS 2024 · similarity 0.86)
- Inverse Reinforcement Learning by Estimating Expertise of Demonstrators (AAAI 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.