エネルギーベースの自己教師あり行動ゲーティングによる拡散プランナーの改善
Improving Diffusion Planners by Self-Supervised Action Gating with Energies
拡散プランナーの推論時に、JEPAで学習した潜在予測誤差をエネルギーとして候補軌道を再ランキングし、動的に矛盾する計画を排除して性能と堅牢性を高める手法SAGEを提案。
詳しい要約
1. どんなもの?
2. 先行研究と比べてどこがすごい?
3. 技術・手法の肝は?
4. どうやって有効だと検証した?
5. 議論はある?
6. 次に読むべき論文は?
※ AIが要旨から生成した要約です。正確性は原文をご確認ください。
著者: Yuan Lu, Dongqi Han, Yansen Wang, Dongsheng Li
分類: cs.LG, cs.AI, cs.RO
原文アブストラクト
Diffusion planners are a strong approach for offline reinforcement learning, but they can fail when value-guided selection favours trajectories that score well yet are locally inconsistent with the environment dynamics, resulting in brittle execution. We propose Self-supervised Action Gating with Energies (SAGE), an inference-time re-ranking method that penalises dynamically inconsistent plans using a latent consistency signal. SAGE trains a Joint-Embedding Predictive Architecture (JEPA) encoder on offline state sequences and an action-conditioned latent predictor for short horizon transitions. At test time, SAGE assigns each sampled candidate an energy given by its latent prediction error and combines this feasibility score with value estimates to select actions. SAGE can integrate into existing diffusion planning pipelines that can sample trajectories and select actions via value scoring; it requires no environment rollouts and no policy re-training. Across locomotion, navigation, and manipulation benchmarks, SAGE improves the performance and robustness of diffusion planners.
