Research map / Reinforcement learning / Visual and representation RL
Transfer of Deep Reactive Policies for MDP Planning
Aniket (Nick) Bajpai, Sankalp Garg, Mausam
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Add action decoder and transition transfer to A3C+GCN (ablation) | TP · Nine reported SysAdmin, Game of Life, and Navigation instances | α (zero-shot): Sys#1: 0.01 -> 0.25; Sys#5: 0.01 -> 0.23; Sys#10: 0.01 -> 0.22; Game#1: 0.03 -> 0.31; Game#5: 0.02 -> 0.26; Game#10: 0.0 | Improved |
| Add instance classifier to A3C+GCN+SAD (ablation) | TP · Nine reported SysAdmin, Game of Life, and Navigation instances | α (zero-shot): Sys#1: 0.25 -> 0.23; Sys#5: 0.23 -> 0.26; Sys#10: 0.22 -> 0.26; Game#1: 0.31 -> 0.34; Game#5: 0.26 -> 0.41; Game#10: 0.2 | Mixed |
| Transfer policy and embedding components from source instances and retrain the target decoder | TP · Nine reported SysAdmin, Game of Life, and Navigation instances | α (zero-shot): Sys#1: A2T 0.08, TP 0.23; Sys#5: A2T 0.02, TP 0.26; Sys#10: A2T 0.06, TP 0.26; Game#1: A2T 0.19, TP 0.34; Game#5: A2T 0. | Improved |
| Use TP rather than A2T | TP · Nine reported SysAdmin, Game of Life, and Navigation instances | α (∞ training): TP 1.0 versus A2T 0.38, 0.40, 0.33, 0.61, 0.44, 0.20, 0.25, 0.56, 0.40 | Improved |
Closest papers by meaning
- Deep Reactive Policies for Planning in Stochastic Nonlinear Domains (AAAI 2019 · similarity 0.87)
- Learning Planning Domains from Non-redundant Fully-Observed Traces: Theoretical Foundations and Complexity Analysis (AAAI 2024 · similarity 0.83)
- Model-Based Transfer Learning for Contextual Reinforcement Learning (NeurIPS 2024 · similarity 0.83)
- Diffusion Policy Policy Optimization (ICLR 2025 · similarity 0.83)
- Distilling Reinforcement Learning Algorithms for In-Context Model-Based Planning (ICLR 2025 · similarity 0.83)
- Are Expressive Models Truly Necessary for Offline RL? (AAAI 2025 · similarity 0.82)
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.
