2021/02/14 by Urmish Thakker, Paul N. Whatmough, Thakker, Urmish +7
Computer Science · #Advanced Neural Network Applications #Algorithm #Computer science #FOS: Computer and information sciences #Kronecker delta #Kronecker product #Machine Learning (cs.LG) #Machine Learning and ELM #Materials science #Matrix (chemical analysis) #Matrix decomposition #Matrix multiplication #Sparse matrix #Topic Modeling #cs.LG
paper · pdf · doi:10.48550/arxiv.2102.07071
Accepted to be published at MLSys 2021
arxiv created 2021/02/14 · openalex publication_date 2021/02/14 · arxiv updated 2021/02/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Structured matrices, such as those derived from Kronecker products (KP), are effective at compressing neural networks, but can lead to unacceptable accuracy loss when applied to large models. In this paper, we propose the notion of doping -- addition of an extremely sparse matrix to a structured matrix. Doping facilitates additional degrees of freedom for a small number of parameters, allowing them to independently diverge from the fixed structure. To train LSTMs with doped structured matrices, we introduce the additional parameter matrix while slowly annealing its sparsity level. However, we find that performance degrades as we slowly sparsify the doping matrix, due to co-matrix adaptation (CMA) between the structured and the sparse matrices. We address this over dependence on the sparse matrix using a co-matrix dropout regularization (CMR) scheme. We provide empirical evidence to show that doping, CMA and CMR are concepts generally applicable to multiple structured matrices (Kronecker Product, LMF, Hybrid Matrix Decomposition). Additionally, results with doped kronecker product matrices demonstrate state-of-the-art accuracy at large compression factors (10 - 25x) across 4 natural language processing applications with minor loss in accuracy. Doped KP compression technique outperforms previous state-of-the art compression results by achieving 1.3 - 2.4x higher compression factor at a similar accuracy, while also beating strong alternatives like pruning and low-rank methods by a large margin (8% or more). Additionally, we show that doped KP can be deployed on commodity hardware using the current software stack and achieve 2.5 - 5.5x inference run-time speed-up over baseline.