AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Machine Learning and Overfitting
The idea of overfitting has been around for at least 25, 30, 40 years. It is only in the last 10 years when a technique called dropout was invented by Professor Hinton among others where you are deliberately punching holes in the neural network to reduce overfitting. The performance out of sample performance suddenly had become horrible and the performance of business and commercial problem actually become adequate. Most of the advances in machine learning in the last 20 years is focused on overcoming this particular problem.