Timezone: »

 
Poster
Einconv: Exploring Unexplored Tensor Network Decompositions for Convolutional Neural Networks
Kohei Hayashi · Taiki Yamaguchi · Yohei Sugawara · Shin-ichi Maeda

Thu Dec 12 10:45 AM -- 12:45 PM (PST) @ East Exhibition Hall B + C #139
in Deep Learning -- CNN Architectures »

Tensor decomposition methods are widely used for model compression and fast inference in convolutional neural networks (CNNs). Although many decompositions are conceivable, only CP decomposition and a few others have been applied in practice, and no extensive comparisons have been made between available methods. Previous studies have not determined how many decompositions are available, nor which of them is optimal. In this study, we first characterize a decomposition class specific to CNNs by adopting a flexible graphical notation. The class includes such well-known CNN modules as depthwise separable convolution layers and bottleneck layers, but also previously unknown modules with nonlinear activations. We also experimentally compare the tradeoff between prediction accuracy and time/space complexity for modules found by enumerating all possible decompositions, or by using a neural architecture search. We find some nonlinear decompositions outperform existing ones.

Keywords: Matrix and Tensor Factorization · CNN Architectures · Applications · Deep Learning
[ Paper]  [ Slides] 

Author Information

Kohei Hayashi (Preferred Networks)
Taiki Yamaguchi (The University of Tokyo)
Yohei Sugawara (Preferred Networks, Inc.)
Shin-ichi Maeda (Preferred Networks)

More from the Same Authors