Week 2 - 13th & 16th June
12 Jun 2017Learning ReLUs via gradient descent
In our 3rd meet last Tuesday, Gauri volunteered to present the paper titled Learning ReLUs via gradient descent – theoretical results proving linear convergence of projected gradient descent algorithm, to fit Rectified Linear Units (ReLUs) to data, using optimal number of samples.
Notes for Tuesday meeting are here.
In next meet on Friday, Gauri concluded the topic of Tuesday’s meet.
Notes for Friday meeting will be uploaded once available.