Shivon Zilis@shivon·Apr 1, 2015Great read! From linear models to neural networks from @Reza_Zadeh and @dbeyer123 http://amplifypartners.com/interviews/on-the-evolution-of-machine-learning-from-linear-models-to-neural-networks/…1819
Matt Gershoff@mgershoffReplying to @shivon@shivon @Reza_Zadeh Isn't the main trade off losing convexity. I think that is what you are hinting at.3:48 AM · Apr 1, 2015·Twitter Web Client
Reza Zadeh@Reza_Zadeh·Apr 1, 2015Replying to @mgershoff@mgershoff @shivon Yes, that's what I was hinting at.1
Matt Gershoff@mgershoff·Apr 1, 2015@Reza_Zadeh @shivon The sticky part is adaptive features, we can still have a nonlinear map from feature/Kernel space with a linear model
