Research map / Reinforcement learning / Imitation learning
Machine Teaching for Inverse Reinforcement Learning: Algorithms and Applications
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 SCOT instead of UVM with 10^7 Monte Carlo samples | IRL demonstration teaching · Random 9x9 grid worlds, 8-dimensional binary features | Avg. policy loss: 0.555 -> 0.001 | Helped |
| Use SCOT instead of UVM with 10^7 Monte Carlo samples | IRL demonstration teaching · Random 9x9 grid worlds, 8-dimensional binary features | Avg. time (s): 10291.365 -> 0.965 | Helped |
| Use SCOT instead of random demonstrations | IRL demonstration teaching · Random 9x9 grid worlds, 8-dimensional binary features | Ave. policy loss: 0.015 -> 0.001 | Helped |
| Remove redundant half-space constraints during SCOT (ablation) | SCOT · Random 9x9 grid worlds, 8-dimensional binary features | Ave. (s,a) pairs: 66.740 -> 17.160 | Helped |
| Remove redundant half-space constraints during SCOT (ablation) | SCOT · Random 9x9 grid worlds, 8-dimensional binary features | Ave. time (s): 12.407 -> 0.965 | Helped |
| Use UVM with more Monte Carlo samples | UVM · Random 9x9 grid worlds, 8-dimensional binary features | Avg. time (s): 247.944 -> 10291.365 | Hurt |
| Use BIO-IRL's informative-demonstration likelihood instead of standard BIRL likelihood | Bayesian IRL · Ball sorting task | policy value loss | Helped |
| Use BIO-IRL's informative-demonstration likelihood instead of standard BIRL likelihood | Bayesian IRL · Markov chain example | posterior concentration | Helped |
| Increase UVM Monte Carlo samples from 10^5 to 10^7 | UVM · Random 9x9 grid worlds, 8-dimensional binary features | Avg. policy loss: 1.539 -> 0.555 | Helped |
| Increase UVM Monte Carlo samples from 10^5 to 10^7 | UVM · Random 9x9 grid worlds, 8-dimensional binary features | Avg. % incorrect actions: 31.420 -> 18.642 | Helped |
Closest papers by meaning
- Teaching Inverse Reinforcement Learners via Features and Demonstrations (NeurIPS 2018 · similarity 0.92)
- Hybrid Inverse Reinforcement Learning (ICML 2024 · similarity 0.90)
- Is Inverse Reinforcement Learning Harder than Standard Reinforcement Learning? A Theoretical Perspective (ICML 2024 · similarity 0.90)
- Recent Advancements in Inverse Reinforcement Learning (AAAI 2024 · similarity 0.90)
- Inverse Reinforcement Learning by Estimating Expertise of Demonstrators (AAAI 2025 · similarity 0.89)
- Offline Inverse RL: New Solution Concepts and Provably Efficient Algorithms (ICML 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.