
…networks. Fundamentally, neural networks are nothing more than really good function approximators — you give a trained network an input vector, it performs a series of operations, and it produces an output vector. To train our network to estimate an unknown function, we give it a collection of data points — whic…
…o Stanford University as an undergrad student. A few months ago, I achieved this goal. At Stanford, I’ll probably be studying Symbolic Systems, which is a program that explores both the humanities and STEM to inform an understanding of artificial intelligence and the nature of minds. Needless to say, A Year of AI will continue to document the new things I learn 😀.