Topics
Explore 4 articles on Research updates from How Robots Learn.
TERRA learns latent actions from compact representations of temporal visual change. Its EAT mechanism aligns the effects produced by the same code across different initial states, while the reported evaluations remain claims from the preprint’s authors.
Written
RLHND converts a pretrained Cosmos 3 video diffusion backbone into a deterministic clip-level feature extractor for monocular egocentric recordings. Separate streams estimate anatomically plausible hand pose and dense surface contact and force; the authors also claim leading benchmark performance and robot-learning utility, but the available abstract contains no numerical or task-level evidence.
Written
Bee uses the consistency of human corrections to constrain a robot policy, with results from three physical tasks and one simulation.
Written
How RoboRecover reconstructs simulated recovery points, compares clean-start and recovery results on LIBERO, and limits what the rates can show.
Written