Published onOctober 22, 2025MIT’s SoftMimic Framework Teaches Humanoids to Be Compliant, Not Just CapableResearchers from MIT's Improbable AI Lab have developed SoftMimic, a reinforcement learning framework that trains humanoids to be compliant and safely absorb collisions, rather than rigidly tracking motions.MITReinforcement-LearningSoftMimicRead more →