DNN gradient lossless compression: Can GenNorm be the answer?

Zhong Jing Chen, Eduin E. Hernandez, Yu-Chih Huang, Stefano Rini

研究成果: Conference contribution同行評審

摘要

In this paper, the problem of optimal gradient lossless compression in Deep Neural Network (DNN) training is considered. Gradient compression is relevant in many distributed DNN training scenarios, including the recently popular federated learning (FL) scenario in which each remote users are connected to the parameter server (PS) through a noiseless but rate limited channel. In distributed DNN training, if the underlying gradient distribution is available, classical lossless compression approaches can be used to reduce the number of bits required for communicating the gradient entries. Mean field analysis has suggested that gradient updates can be considered as independent random variables, while Laplace approximation can be used to argue that gradient has a distribution approximating the normal (Norm) distribution in some regimes. In this paper we argue that, for some networks of practical interest, the gradient entries can be well modelled as having a generalized normal (GenNorm) distribution. We provide numerical evaluations to validate that the hypothesis GenNorm modelling provides a more accurate prediction of the DNN gradient tail distribution. Additionally, this modeling choice provides concrete improvement in terms of lossless compression of the gradients when applying classical fix-to-variable lossless coding algorithms, such as Huffman coding, to the quantized gradient updates. This latter results indeed provides an effective compression strategy with low memory and computational complexity that has great practical relevance in distributed DNN training scenarios.

原文English
主出版物標題ICC 2022 - IEEE International Conference on Communications
發行者Institute of Electrical and Electronics Engineers Inc.
頁面407-412
頁數6
ISBN(電子)9781538683477
DOIs
出版狀態Published - 2022
事件2022 IEEE International Conference on Communications, ICC 2022 - Seoul, Korea, Republic of
持續時間: 16 5月 202220 5月 2022

出版系列

名字IEEE International Conference on Communications
2022-May
ISSN(列印)1550-3607

Conference

Conference2022 IEEE International Conference on Communications, ICC 2022
國家/地區Korea, Republic of
城市Seoul
期間16/05/2220/05/22

指紋

深入研究「DNN gradient lossless compression: Can GenNorm be the answer?」主題。共同形成了獨特的指紋。

引用此