Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

Picture for Philip M. Long

Properties of the After Kernel


May 27, 2021
Philip M. Long


  Access Paper or Ask Questions

When does gradient descent with logistic loss interpolate using deep networks with smoothed ReLU activations?


Feb 09, 2021
Niladri S. Chatterji, Philip M. Long, Peter L. Bartlett

* 97 pages 

  Access Paper or Ask Questions

When does gradient descent with logistic loss find interpolating two-layer networks?


Dec 04, 2020
Niladri S. Chatterji, Philip M. Long, Peter L. Bartlett

* 43 pages, 4 figures 

  Access Paper or Ask Questions

Failures of model-dependent generalization bounds for least-norm interpolation


Oct 30, 2020
Peter L. Bartlett, Philip M. Long


  Access Paper or Ask Questions

Finite-sample analysis of interpolating linear classifiers in the overparameterized regime


Apr 25, 2020
Niladri S. Chatterji, Philip M. Long


  Access Paper or Ask Questions

On the Global Convergence of Training Deep Linear ResNets


Mar 02, 2020
Difan Zou, Philip M. Long, Quanquan Gu

* 26 pages, 1 figure. In ICLR 2020 

  Access Paper or Ask Questions

Oracle lower bounds for stochastic gradient sampling algorithms


Feb 01, 2020
Niladri S. Chatterji, Peter L. Bartlett, Philip M. Long

* 21 pages 

  Access Paper or Ask Questions

Benign Overfitting in Linear Regression


Jun 26, 2019
Peter L. Bartlett, Philip M. Long, Gábor Lugosi, Alexander Tsigler


  Access Paper or Ask Questions

Size-free generalization bounds for convolutional neural networks


Jun 12, 2019
Philip M. Long, Hanie Sedghi


  Access Paper or Ask Questions

On the effect of the activation function on the distribution of hidden nodes in a deep network


Jan 07, 2019
Philip M. Long, Hanie Sedghi


  Access Paper or Ask Questions

Density estimation for shift-invariant multidimensional distributions


Nov 09, 2018
Anindya De, Philip M. Long, Rocco A. Servedio

* Appears in the Proceedings of ITCS 2019 

  Access Paper or Ask Questions

Learning Sums of Independent Random Variables with Sparse Collective Support


Jul 18, 2018
Anindya De, Philip M. Long, Rocco Servedio

* Conference version appears in Proceedings of FOCS 2018 

  Access Paper or Ask Questions

Gradient descent with identity initialization efficiently learns positive definite linear transformations by deep residual networks


Jun 18, 2018
Peter L. Bartlett, David P. Helmbold, Philip M. Long


  Access Paper or Ask Questions

The Power of Localization for Efficiently Learning Linear Separators with Noise


Jun 03, 2018
Pranjal Awasthi, Maria Florina Balcan, Philip M. Long

* Contains improved label complexity analysis communicated to us by Steve Hanneke 

  Access Paper or Ask Questions

The Singular Values of Convolutional Layers


May 26, 2018
Hanie Sedghi, Vineet Gupta, Philip M. Long


  Access Paper or Ask Questions

Representing smooth functions as compositions of near-identity functions with implications for deep network optimization


Apr 16, 2018
Peter L. Bartlett, Steven N. Evans, Philip M. Long


  Access Paper or Ask Questions

Surprising properties of dropout in deep networks


Apr 19, 2017
David P. Helmbold, Philip M. Long


  Access Paper or Ask Questions

On the Inductive Bias of Dropout


Feb 17, 2015
David P. Helmbold, Philip M. Long

* Journal of Machine Learning Research, 16, 3403-3454 (2015). (See http://jmlr.org/papers/volume16/helmbold15a/helmbold15a.pdf.) 

  Access Paper or Ask Questions

Active and passive learning of linear separators under log-concave distributions


Apr 26, 2013
Maria Florina Balcan, Philip M. Long


  Access Paper or Ask Questions

On the Necessity of Irrelevant Variables


Jun 08, 2012
David P. Helmbold, Philip M. Long

* A preliminary version of this paper appeared in the proceedings of ICML'11 

  Access Paper or Ask Questions