Challenges and opportunities toward online training acceleration using RRAM-based hardware neural network
Authors:
- 1 December 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 11.6.1-11.6.4
Abstract
This paper highlights the feasible routes of using resistive memory (RRAM) for accelerating online training of deep neural networks (DNNs). A high degree of asymmetric nonlinearity in analog RRAMs could be tolerated when weight update algorithms are optimized with reduced training noise. Hybrid-weight Net (HW-Net), a modified multilayer perceptron (MLP) algorithm that utilizes hybrid internal analog and external binary weights is also proposed. Highly accurate online training could be realized using simple binary RRAMs that have already been widely developed as digital memory.Keywords
This publication has 13 references indexed in Scilit:
- Binary neural network with 16 Mb RRAM macro chip for classification and online training2016
- Resistive memory device requirements for a neural algorithm accelerator2016
- Dot-product engine for neuromorphic computing2016
- Characterization and Modeling of Nonfilamentary Ta/TaOx/TiO2/Ti Analog Synaptic DeviceScientific Reports, 2015
- Training and operation of an integrated neuromorphic network based on metal-oxide memristorsNature, 2015
- Variability-tolerant Convolutional Neural Network for Pattern Recognition applications based on OxRAM synapses2014
- Experimental demonstration and tolerancing of a large-scale neural network (165,000 synapses), using phase-change memory as the synaptic weight element2014
- 3D synaptic architecture with ultralow sub-10 fJ energy per spike for neuromorphic computation2014
- Statistical Model and Rapid Prediction of RRAM SET Speed–Disturb DilemmaIEEE Transactions on Electron Devices, 2013
- Analysis and modeling of resistive switching statisticsJournal of Applied Physics, 2012