A more effective way to train machines for uncertain, real-w

A new algorithm developed at MIT determines whether a machine-learning system should try to mimic its teacher or explore on its own through trial-and-error.

Related Keywords

Israel , Rishabh Agarwal , Idan Shenfeld , Zhang Wei Hong , Pulkit Agrawal , Abhishek Gupta , Aviv Tamar , Hyundai Motor Company , Israel Institute Of Technology , Improbable Ai Lab , Office Of Naval Research , Machine Common Sense Program , University Of Washington , Artificial Intelligence Laboratory , International Conference On Machine , Watson Ai Lab , Israel Institute , Computer Science , Artificial Intelligence , International Conference , Google Brain , Common Sense Program , Eacher Guide Reinforcement Learning , Eacher Student Computational Learning ,