Imitation Learning on General MDPs various settings
1Expert Trajectories (τ_E)ILARL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ILARLSetting=Linear MDP, W_{max} = 1, F.O. (Features Only)=✓2025.02 | 1 | 3 | |
| Behavioural CloningSetting=Function Approximation, Episodic Foster et al. (2024), F.O. (Features Only)=✗2025.02 | 2 | — | |
| Behavioural CloningSetting=Tabular, Episodic Rajaraman et al. (2020), F.O. (Features Only)=✗2025.02 | 2 | — | |
| Behavioural CloningSetting=Deterministic Linear Expert, Episodic Rajaraman et al. (2021), F.O. (Features Only)=✗2025.02 | 2 | — | |
| OALSetting=Episodic Tabular, F.O. (Features Only)=✗2025.02 | 2 | 4 | |
| MobileSetting=Episodic, r_{true} ∈ R and P ∈ P, F.O. (Features Only)=✓⋆2025.02 | 2 | 5 | |
| FRA-ILSetting=Linear MDP, F.O. (Features Only)=✓2025.02 | 2 | 3 | |
| Lower BoundSetting=Linear MDP, F.O. (Features Only)=✓2025.02 | 2 | 1 | |
| Mimic-MDSetting=Tabular, Known P, Deterministic Expert, Episodic, F.O. (Features Only)=✗2025.02 | 3 | — | |
| MB-TAILSetting=Episodic, Tabular, Deterministic Expert, F.O. (Features Only)=✗2025.02 | 3 | 3 | |
| OGAILSetting=Episodic Linear Mixture MDP, W_{max} = √d, F.O. (Features Only)=✓2025.02 | 3 | 4 | |
| FAILSetting=Episodic, πE ∈ Π and V^{πE} ∈ F, F.O. (Features Only)=✓⋆2025.02 | 4 | 4 |