Practically Relevant Theory of Deep Learning

Project Information

TheoryDL

Grant agreement ID: 676774

DOI

10.3030/676774

Project closed

EC signature date 7 December 2015

Start date 1 February 2016

End date 31 January 2021

Funded under

• EXCELLENT SCIENCE - European Research Council (ERC)

Total cost

€ 1 342 500,00

EU contribution

€ 1 342 500,00

Coordinated by

THE HEBREW UNIVERSITY OF JERUSALEM
Israel

CORDIS provides links to public deliverables and publications of HORIZON projects.

Links to deliverables and publications from FP7 projects, as well as links to some specific result types such as dataset and software, are dynamically retrieved from OpenAIRE .

Publications

Conference proceedings (11)

SDCA without Duality, Regularization, and Individual Convexity

Author(s): Shai Shalev-Shwartz
Published in: ICML, 2016
Publisher: ICML

Failures of Gradient-Based Deep Learning

Author(s): Shai Shalev-Shwartz, Ohad Shamir, Shaked Shammah
Published in: ICML, 2017
Publisher: ICML

"Decoupling ""when to update"" from ""how to update"""

Author(s): Eran Malach, Shai Shalev-Shwartz
Published in: NIPS, 2017
Publisher: Curran Associates, Inc.

Complexity theoretic limitations on learning DNF's

Author(s): Amit Daniely, Shai Shalev-Shwartz.
Published in: COLT, 2016
Publisher: COLT

On Graduated Optimization for Stochastic Non-Convex Problems

Author(s): Elad Hazan, Kfir Y. Levy, Shai Shalev-Shwartz
Published in: ICML, 2016
Publisher: ICML

Fast Rates for Empirical Risk Minimization of Strict Saddle Problems

Author(s): Alon Gonen, Shai Shalev-Shwartz
Published in: COLT, 2017
Publisher: COLT

SGD Learns Over-parameterized Networks that Provably Generalize on Linearly Separable Data

Author(s): Alon Brutzkus, Amir Globerson, Eran Malach, Shai Shalev-Shwartz
Published in: International Conference on Learning Representations (ICLR), 2018
Publisher: ICLR

Solving Ridge Regression using Sketched Preconditioned SVRG

Author(s): Alon Gonen, Francesco Orabona, Shai Shalev-Shwartz
Published in: ICML, 2016
Publisher: ICML

On the Expressive Power of Overlapping Architectures of Deep Learning

Author(s): Or Sharir, Amnon Shashua
Published in: ICLR, 2018
Publisher: ICLR

Sum-Product-Quotient Networks

Author(s): Or Sharir, Amnon Shashua
Published in: AISTATS, 2018
Publisher: AISTATS

On the Long-Term Memory of Deep Recurrent Networks

Author(s): Yoav Levine, Or Sharir, Alon Ziv, Amnon Shashua
Published in: ICLR, 2018
Publisher: ICLR workshop

Other (2)

Weight Sharing is Crucial to Succesful Optimization

Author(s): Shai Shalev-Shwartz, Ohad Shamir, Shaked Shammah
Published in: 2017
Publisher: arxiv

A Provably Correct Algorithm for Deep Learning that Actually Works

Author(s): Eran Malach, Shai Shalev-Shwartz
Published in: 2018
Publisher: arxiv