Collective Experience Learning Method Developed for Robots at Tsinghua University

Researchers from Tsinghua University, Infinigence AI, and AgiBot have proposed a new method called Collective Experience Learning (CEL), where robots learn from each other's successes instead of relying solely on human demonstrations. This method was tested on a cluster of 96 robots, showing remarkable results across five different tasks, such as flower arrangement and stool assembly, achieving an overall success rate of 97.5%, surpassing the average performance of six human operators.