AINeutralarXiv – CS AI · 3h ago6/10
🧠
Visualizing Latent Phase Structures in Locomotion Policies: A Multi-Environment Study with Temporal Feature Extension
Researchers propose a novel framework for visualizing latent motion phase structures in deep reinforcement learning locomotion policies by extending clustering features beyond state observations to include actions and next states. The method successfully identifies clearer phase transition patterns across three MuJoCo environments, advancing interpretability of neural network-based control policies.