## Abstract

Expressive efficiency refers to the relation between two architectures A and B, whereby any function realized by B could be replicated by A, but there exists functions realized by A, which cannot be replicated by B unless its size grows significantly larger. For example, it is known that deep networks are exponentially efficient with respect to shallow networks, in the sense that a shallow network must grow exponentially large in order to approximate the functions represented by a deep network of polynomial size. In this work, we extend the study of expressive efficiency to the attribute of network connectivity and in particular to the effect of "overlaps" in the convolutional process, i.e., when the stride of the convolution is smaller than its filter size (receptive field). To theoretically analyze this aspect of network's design, we focus on a well-established surrogate for ConvNets called Convolutional Arithmetic Circuits (ConvACs), and then demonstrate empirically that our results hold for standard ConvNets as well. Specifically, our analysis shows that having overlapping local receptive fields, and more broadly denser connectivity, results in an exponential increase in the expressive capacity of neural networks. Moreover, while denser connectivity can increase the expressive capacity, we show that the most common types of modern architectures already exhibit exponential increase in expressivity, without relying on fully-connected layers.

Original language | American English |
---|---|

State | Published - 2018 |

Event | 6th International Conference on Learning Representations, ICLR 2018 - Vancouver, Canada Duration: 30 Apr 2018 → 3 May 2018 |

### Conference

Conference | 6th International Conference on Learning Representations, ICLR 2018 |
---|---|

Country/Territory | Canada |

City | Vancouver |

Period | 30/04/18 → 3/05/18 |

### Bibliographical note

Funding Information:This work is supported by Intel grant ICRI-CI #9-2012-6133, by ISF Center grant 1790/12 and by the European Research Council (TheoryDL project).

Publisher Copyright:

© 2017 International Conference on Learning Representations, ICLR. All rights reserved.