# Imitation Learning

Most robot policies output action values at discrete control steps and leave the continuity of physical motion to be learned implicitly from data. ODE Policy instead learns a neural ordinary differential equation for the vector field of the action, and at deployment integrates each action from the previous one, so the executed trajectory is continuous by construction. After one epoch on LIBERO it reaches 79.2% success against at most 43.2% for the baselines, and it keeps that early lead as the action head of a VLA and of ImageWAM and as the actor of SuperMimic.

Most robot policies output action values at discrete control steps and leave the continuity of physical motion to be learned implicitly from data. ODE Policy instead learns a neural ordinary differential equation for the vector field of the action, and at deployment integrates each action from the previous one, so the executed trajectory is continuous by construction. After one epoch on LIBERO it reaches 79.2% success against at most 43.2% for the baselines, and it keeps that early lead as the action head of a VLA and of ImageWAM and as the actor of SuperMimic.

The work identifies that in decentralized multi-agent path finding, even correctly learned per-agent action distributions can be recombined by independent sampling into incompatible joint actions, and proposes DMM, which replaces one-shot sampling with discrete iterative refinement of action intents across communication rounds before commitment, fine-tuned by MICPO, a critic-free multi-agent reinforcement-learning method; DMM generally achieves higher success rates and lower solution costs on POGEMA, solves 1,598 of 1,600 MovingAI tasks (the highest coverage among the evaluated methods), and scales to over one million simultaneously acting agents.

The work identifies that in decentralized multi-agent path finding, even correctly learned per-agent action distributions can be recombined by independent sampling into incompatible joint actions, and proposes DMM, which replaces one-shot sampling with discrete iterative refinement of action intents across communication rounds before commitment, fine-tuned by MICPO, a critic-free multi-agent reinforcement-learning method; DMM generally achieves higher success rates and lower solution costs on POGEMA, solves 1,598 of 1,600 MovingAI tasks (the highest coverage among the evaluated methods), and scales to over one million simultaneously acting agents.

The work formalizes multimodality in behavioral cloning, proves that latent-variable policies preserve demonstrated modes only if the latent carries action-conditioned information and that excessive posterior-prior regularization suppresses it, shows that action-space generative policies are limited by the Lipschitz constant of the base-to-action map, validates these mechanisms on synthetic multimodal navigation and a real-robot bimodal tissue-grasping task, and uses a GMM modal-clustering diagnostic to find limited conditional multimodality in Push-T, UR3, LIBERO, and Meta-World, where deterministic regression remains competitive.

The work formalizes multimodality in behavioral cloning, proves that latent-variable policies preserve demonstrated modes only if the latent carries action-conditioned information and that excessive posterior-prior regularization suppresses it, shows that action-space generative policies are limited by the Lipschitz constant of the base-to-action map, validates these mechanisms on synthetic multimodal navigation and a real-robot bimodal tissue-grasping task, and uses a GMM modal-clustering diagnostic to find limited conditional multimodality in Push-T, UR3, LIBERO, and Meta-World, where deterministic regression remains competitive.