Early-Signal Detector

method · first seen 5 Oct 2026

learning from human video demonstrations

Also: video imitation learning, learning from human videos

Training robot policies using egocentric videos of humans performing tasks rather than robot teleoperation data.

Mentions per week

0319 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

dexterous manipulationdiffusion policyvision-language-action modelhumanoid whole-body controlcode-as-policymobile manipulationGaussian splattingreinforcement learning from world model predictions