1 paper
Sam Dillavou, Jason W Rocks, Jacob F Wycoff +2
An important component of the success of large AI models is double descent, in which networks avoid overfitting as they grow relative to the amount of training data, instead improv…