AdaNet: Adaptive structural learning of artificial neural networks

Corinna Cortes, Xavier Gonzalvo, Vitaly Kuznetsov, Mehryar Mohri, Scott Yang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We present new algorithms for adaptively learning artificial neural networks. Our algorithms (AdaNet) adaptively learn both the structure of the network and its weights. They are based on a solid theoretical analysis, including data-dependent generalization guarantees that we prove and discuss in detail. We report the results of large-scale experiments with one of our algorithms on several binary classification tasks extracted from the CIFAR-10 dataset and on the Criteo dataset. The results demonstrate that our algorithm can automatically learn network structures with very competitive performance accuracies when compared with those achieved by neural networks found by standard approaches.

Original languageEnglish (US)
Title of host publication34th International Conference on Machine Learning, ICML 2017
PublisherInternational Machine Learning Society (IMLS)
Pages1452-1466
Number of pages15
Volume2
ISBN (Electronic)9781510855144
StatePublished - Jan 1 2017
Event34th International Conference on Machine Learning, ICML 2017 - Sydney, Australia
Duration: Aug 6 2017Aug 11 2017

Other

Other34th International Conference on Machine Learning, ICML 2017
CountryAustralia
CitySydney
Period8/6/178/11/17

Fingerprint

Neural networks
Experiments

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Human-Computer Interaction
  • Software

Cite this

Cortes, C., Gonzalvo, X., Kuznetsov, V., Mohri, M., & Yang, S. (2017). AdaNet: Adaptive structural learning of artificial neural networks. In 34th International Conference on Machine Learning, ICML 2017 (Vol. 2, pp. 1452-1466). International Machine Learning Society (IMLS).

AdaNet : Adaptive structural learning of artificial neural networks. / Cortes, Corinna; Gonzalvo, Xavier; Kuznetsov, Vitaly; Mohri, Mehryar; Yang, Scott.

34th International Conference on Machine Learning, ICML 2017. Vol. 2 International Machine Learning Society (IMLS), 2017. p. 1452-1466.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Cortes, C, Gonzalvo, X, Kuznetsov, V, Mohri, M & Yang, S 2017, AdaNet: Adaptive structural learning of artificial neural networks. in 34th International Conference on Machine Learning, ICML 2017. vol. 2, International Machine Learning Society (IMLS), pp. 1452-1466, 34th International Conference on Machine Learning, ICML 2017, Sydney, Australia, 8/6/17.
Cortes C, Gonzalvo X, Kuznetsov V, Mohri M, Yang S. AdaNet: Adaptive structural learning of artificial neural networks. In 34th International Conference on Machine Learning, ICML 2017. Vol. 2. International Machine Learning Society (IMLS). 2017. p. 1452-1466
Cortes, Corinna ; Gonzalvo, Xavier ; Kuznetsov, Vitaly ; Mohri, Mehryar ; Yang, Scott. / AdaNet : Adaptive structural learning of artificial neural networks. 34th International Conference on Machine Learning, ICML 2017. Vol. 2 International Machine Learning Society (IMLS), 2017. pp. 1452-1466
@inproceedings{f2556c0a7a8442718214e835050730ac,
title = "AdaNet: Adaptive structural learning of artificial neural networks",
abstract = "We present new algorithms for adaptively learning artificial neural networks. Our algorithms (AdaNet) adaptively learn both the structure of the network and its weights. They are based on a solid theoretical analysis, including data-dependent generalization guarantees that we prove and discuss in detail. We report the results of large-scale experiments with one of our algorithms on several binary classification tasks extracted from the CIFAR-10 dataset and on the Criteo dataset. The results demonstrate that our algorithm can automatically learn network structures with very competitive performance accuracies when compared with those achieved by neural networks found by standard approaches.",
author = "Corinna Cortes and Xavier Gonzalvo and Vitaly Kuznetsov and Mehryar Mohri and Scott Yang",
year = "2017",
month = "1",
day = "1",
language = "English (US)",
volume = "2",
pages = "1452--1466",
booktitle = "34th International Conference on Machine Learning, ICML 2017",
publisher = "International Machine Learning Society (IMLS)",

}

TY - GEN

T1 - AdaNet

T2 - Adaptive structural learning of artificial neural networks

AU - Cortes, Corinna

AU - Gonzalvo, Xavier

AU - Kuznetsov, Vitaly

AU - Mohri, Mehryar

AU - Yang, Scott

PY - 2017/1/1

Y1 - 2017/1/1

N2 - We present new algorithms for adaptively learning artificial neural networks. Our algorithms (AdaNet) adaptively learn both the structure of the network and its weights. They are based on a solid theoretical analysis, including data-dependent generalization guarantees that we prove and discuss in detail. We report the results of large-scale experiments with one of our algorithms on several binary classification tasks extracted from the CIFAR-10 dataset and on the Criteo dataset. The results demonstrate that our algorithm can automatically learn network structures with very competitive performance accuracies when compared with those achieved by neural networks found by standard approaches.

AB - We present new algorithms for adaptively learning artificial neural networks. Our algorithms (AdaNet) adaptively learn both the structure of the network and its weights. They are based on a solid theoretical analysis, including data-dependent generalization guarantees that we prove and discuss in detail. We report the results of large-scale experiments with one of our algorithms on several binary classification tasks extracted from the CIFAR-10 dataset and on the Criteo dataset. The results demonstrate that our algorithm can automatically learn network structures with very competitive performance accuracies when compared with those achieved by neural networks found by standard approaches.

UR - http://www.scopus.com/inward/record.url?scp=85048410351&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048410351&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85048410351

VL - 2

SP - 1452

EP - 1466

BT - 34th International Conference on Machine Learning, ICML 2017

PB - International Machine Learning Society (IMLS)

ER -