A Circuit Implementation Method for Memristor Crossbar with On-chip Training

This paper studies circuit realization of artificial neural network (ANN) implemented in a memristor crossbar. We propose an on-chip training circuit by adding an extra memristor column for the storage of negative gradient. With proper control circuit, the proposed circuit can perform both backward and forward signal propagations. For demonstrate we have written a Verilog-AMS based implementation of multi-layer perceptron to validate the proposed design. In addition to reporting simulation results, we also point out the AMS-based simulation bottleneck that requires further investigation.

[1]  Chris Yakopcic,et al.  On-chip training of memristor crossbar based multi-layer neural networks , 2017, Microelectron. J..

[2]  Qin Li,et al.  High linearity source-follower buffer based analog memory for analog convolutional neural network , 2018, Microelectron. J..

[3]  Witali L. Dunin-Barkowski,et al.  An approximate backpropagation learning rule for memristor based neural networks using synaptic plasticity , 2015, Neurocomputing.

[4]  Fabien Alibart,et al.  Pattern classification by memristive crossbar circuits using ex situ and in situ training , 2013, Nature Communications.

[5]  Chris Yakopcic,et al.  On-chip training of memristor based deep neural networks , 2017, 2017 International Joint Conference on Neural Networks (IJCNN).

[6]  Majid Ahmadi,et al.  Analog Implementation of a Novel Resistive-Type Sigmoidal Neuron , 2012, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[7]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[8]  Bo Li,et al.  A current-feedback method for programming memristor array in bidirectional associative memory , 2017, 2017 International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS).

[9]  Yu Wang,et al.  Training itself: Mixed-signal training acceleration for memristor-based neural network , 2014, 2014 19th Asia and South Pacific Design Automation Conference (ASP-DAC).

[10]  Farnood Merrikh-Bayat,et al.  Training and operation of an integrated neuromorphic network based on metal-oxide memristors , 2014, Nature.

[11]  Wei Yang Lu,et al.  Nanoscale memristor device as synapse in neuromorphic systems. , 2010, Nano letters.

[12]  L. Chua.  Memristor-The missing circuit element , 1971 .