Research map / Label-efficient and robust learning / Domain adaptation and generalization
Cross-Domain Policy Optimization via Bellman Consistency and Hybrid Critics
Ming-Hong Chen, Kuan-Chen Pan, You-De Huang, Xi Liu, Ping-Chun Hsieh
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 |
|---|---|---|---|
| transfer source-domain Q via hybrid critic with cross-domain Bellman mapping | SAC · locomotion, Robosuite arm manipulation, Safety-Gym Car->Doggo | data efficiency: about 44% of SAC's environment steps to threshold in best case | Improved |
| use adaptive alpha(t) weighting with a low-quality source model (ablation) | QAvatar · Cheetah with source return 1000 | negative transfer: alpha(t) decreases to a small value by end of training | Improved |
| fix mapping to identity (adversarial source Q) versus learning mapping (ablation) | QAvatar · Ant backward-goal transfer | alpha(t) / transfer | Mixed |
| vary alpha update interval N_alpha = 300, 1000, 3000 (ablation) | QAvatar · Cheetah and Table Wiping | performance | No clear effect |
| transfer from unrelated source (Hopper to table wiping) | QAvatar · Hopper->Robosuite table wiping | negative transfer | No clear effect |
| add second source model (multi-source) (ablation) | QAvatar · Ant, two sources with disabled legs | transferability | Improved |
| fine-tune source models directly instead of QAvatar | FT baseline · Robosuite | learning speed | Hurt |
Closest papers by meaning
- Cross-Domain Policy Adaptation by Capturing Representation Mismatch (ICML 2024 · similarity 0.89)
- Revisiting Reinforcement Learning for LLM Reasoning from A Cross-Domain Perspective (NeurIPS 2025 · similarity 0.87)
- Can RLHF be More Efficient with Imperfect Reward Models? A Policy Coverage Perspective (ICML 2025 · similarity 0.87)
- In-Context Policy Adaptation via Cross-Domain Skill Diffusion (AAAI 2025 · similarity 0.87)
- Dual RL: Unification and New Methods for Reinforcement and Imitation Learning (ICLR 2024 · similarity 0.87)
- Contrastive Representation for Data Filtering in Cross-Domain Offline Reinforcement Learning (ICML 2024 · similarity 0.87)
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.
