Contrastive Energy Fields for Inference-Time Procedure Planning in Instructional Videos

Excerpt

arXiv:2608.16457v1 Announce Type: cross Abstract: Procedure planning seeks to estimate a sequence of actions to transition from an observed initial state to a given goal state. Current procedure planning approaches directly predict action sequences from latent representations using feed-forward neural networks or diffusion-based inference. These paradigms treat every action as plausible, lacking the ability to enforce task-specific logical constraints that render certain actions irrelevant or no