On the number of response regions of deep feed forward networks with piece-wise linear activations

preprint2014arXivOpen access

Abstract

This paper explores the complexity of deep feedforward networks with linear pre-synaptic couplings and rectified linear activations. This is a contribution to the growing body of work contrasting the representational power of deep and shallow network architectures. In particular, we offer a framework for comparing deep and shallow models that belong to the family of piecewise linear functions based on computational geometry. We look at a deep rectifier multi-layer perceptron (MLP) with linear outputs units and compare it with a single layer version of the model. In the asymptotic regime, when the number of inputs stays constant, if the shallow model has knkn hidden units and n0n_0 inputs, then the number of linear regions is O(kn0nn0)O(k^{n_0}n^{n_0}). For a kk layer model with nn hidden units on each layer it is Ω(⌊n/n0⌋k−1nn0)Ω(\left\lfloor {n}/{n_0}\right\rfloor^{k-1}n^{n_0}). The number ⌊n/n0⌋k−1\left\lfloor{n}/{n_0}\right\rfloor^{k-1} grows faster than kn0k^{n_0} when nn tends to infinity or when kk tends to infinity and n≥2n0n \geq 2n_0. Additionally, even when kk is small, if we restrict nn to be 2n02n_0, we can show that a deep model has considerably more linear regions that a shallow one. We consider this as a first step towards understanding the complexity of these models and specifically towards providing suitable mathematical tools for future analysis.

Explore this paper’s authors, topics and related work.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Reviews 0

Write a reviewWrite

No reviews yet.

Discussion 0

Add a commentWrite

No comments yet.