Browsing Publications by Author "Hidary, Jack"

Now showing items 1-3 of 3

    • Classical generalization bounds are surprisingly tight for Deep Networks 

      Liao, Qianli; Miranda, Brando; Hidary, Jack; Poggio, Tomaso (Center for Brains, Minds and Machines (CBMM), 2018-07-11)
      Deep networks are usually trained and tested in a regime in which the training classification error is not a good predictor of the test error. Thus the consensus has been that generalization, defined as convergence of the ...

    • Theory IIIb: Generalization in Deep Networks 

      Poggio, Tomaso; Liao, Qianli; Miranda, Brando; Burbanski, Andrzej; Hidary, Jack (Center for Brains, Minds and Machines (CBMM), arXiv.org, 2018-06-29)
      The general features of the optimization problem for the case of overparametrized nonlinear networks have been clear for a while: SGD selects with high probability global minima vs local minima. In the overparametrized ...

    • Theory of Deep Learning III: explaining the non-overfitting puzzle 

      Poggio, Tomaso; Kawaguchi, Kenji; Liao, Qianli; Miranda, Brando; Rosasco, Lorenzo; e.a. (arXiv, 2017-12-30)
      THIS MEMO IS REPLACED BY CBMM MEMO 90 A main puzzle of deep networks revolves around the absence of overfitting despite overparametrization and despite the large capacity demonstrated by zero training error on randomly ...