  


Bounding and Counting Linear Regions of Deep Neural Networks
Thiago Serra(tserragmail.com) Abstract: In this paper, we study the representational power of deep neural networks (DNN) that belong to the family of piecewiselinear (PWL) functions, based on PWL activation units such as rectifier or maxout. We investigate the complexity of such networks by studying the number of linear regions of the PWL function. Typically, a PWL function from a DNN can be seen as a large family of linear functions acting on millions of such regions. We directly build upon the work of Montufar et al. (2014), Montufar (2017) and Raghu et al. (2017) by refining the upper and lower bounds on the number of linear regions for rectified and maxout networks. In addition to achieving tighter bounds, we also develop a novel method to perform exact enumeration or counting of the number of linear regions with a mixedinteger linear formulation that maps the input space to output. We use this new capability to visualize how the number of linear regions change while training DNNs. Keywords: deep learning, linear regions, piecewiselinear activations, mixedinteger linear programming, solution couting Category 1: Integer Programming ((Mixed) Integer Linear Programming ) Category 2: Applications  Science and Engineering (Statistics ) Citation: Download: [PDF] Entry Submitted: 01/08/2018 Modify/Update this entry  
Visitors  Authors  More about us  Links  
Subscribe, Unsubscribe Digest Archive Search, Browse the Repository

Submit Update Policies 
Coordinator's Board Classification Scheme Credits Give us feedback 
Optimization Journals, Sites, Societies  