Achieving Lossless Accuracy with Lossy Programming for Efficient Neural-Network Training on NVM-Based Systems

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

1 Scopus Citations
View graph of relations

Author(s)

  • Wei-Chen WANG
  • Yuan-Hao CHANG
  • Chien-Chung HO
  • Yu-Ming CHANG
  • Hung-Sheng CHANG

Related Research Unit(s)

Detail(s)

Original languageEnglish
Article number68
Journal / PublicationACM Transactions on Embedded Computing Systems
Volume18
Issue number5s
Online published19 Oct 2019
Publication statusPublished - Oct 2019

Abstract

Neural networks over conventional computing platforms are heavily restricted by the data volume and performance concerns. While non-volatile memory offers potential solutions to data volume issues, challenges must be faced over performance issues, especially with asymmetric read and write performance. Beside that, critical concerns over endurance must also be resolved before non-volatile memory could be used in reality for neural networks. This work addresses the performance and endurance concerns altogether by proposing a data-aware programming scheme. We propose to consider neural network training jointly with respect to the data-flow and data-content points of view. In particular, methodologies with approximate results over Dual-SET operations were presented. Encouraging results were observed through a series of experiments, where great efficiency and lifetime enhancement is seen without sacrificing the result accuracy.

Research Area(s)

  • Lossy programming, Neural network, Non-volatile memory, Phase-change memory

Citation Format(s)

Achieving Lossless Accuracy with Lossy Programming for Efficient Neural-Network Training on NVM-Based Systems. / WANG, Wei-Chen; CHANG, Yuan-Hao; KUO, Tei-Wei; HO, Chien-Chung; CHANG, Yu-Ming; CHANG, Hung-Sheng.

In: ACM Transactions on Embedded Computing Systems, Vol. 18, No. 5s, 68, 10.2019.

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review