Compressing deep neural networks using a rank-constrained topology

Citations Over TimeTop 10% of 2015 papers

Abstract

We present a general approach to reduce the size of feedforward deep neural networks (DNNs). We propose a rankconstrained topology, which factors the weights in the input layer of the DNN in terms of a low-rank representation: unlike previous work, our technique is applied at the level of the filters learned at individual hidden layer nodes, and exploits the natural two-dimensional time-frequency structure in the input. These techniques are applied on a small-footprint DNN-based keyword spotting task, where we find that we can reduce model size by 75% relative to the baseline, without any loss in performance. Furthermore, we find that the proposed approach is more effective at improving model performance compared to other popular dimensionality reduction techniques, when evaluated with a comparable number of parameters.

Related Papers

• → Improving networks using group-based topologies(2008)27 cited
• → Generating realistic ISP-level network topologies(2003)27 cited
• → A survey on topologies of multilevel converters and study of two novel topologies(2009)24 cited
• → When does cp-rank equal rank?(2013)1 cited
• Basic overconstrained topologies with schönflies motions(2014)