Memristor Bridge Synapse-Based Neural Network and Its Learning

Abstract

Analog hardware architecture of a memristor bridge synapse-based multilayer neural network and its learning scheme is proposed. The use of memristor bridge synapse in the proposed architecture solves one of the major problems, regarding nonvolatile weight storage in analog neural network implementations. To compensate for the spatial nonuniformity and nonideal response of the memristor bridge synapse, a modified chip-in-the-loop learning scheme suitable for the proposed neural network architecture is also proposed. In the proposed method, the initial learning is conducted in software, and the behavior of the software-trained network is learned by the hardware network by learning each of the single-layered neurons of the network independently. The forward calculation of the single-layered neuron learning is implemented on circuit hardware, and followed by a weight updating phase assisted by a host computer. Unlike conventional chip-in-the-loop learning, the need for the readout of synaptic weights for calculating weight updates in each epoch is eliminated by virtue of the memristor bridge synapse and the proposed learning scheme. The hardware architecture along with the successful implementation of proposed learning on a three-bit parity network, and on a car detection network is also presented.

A study on emerging electronics for systems accepting soft errors

“…The output of each column can be sent to a amplifier that integrates and fires back to the corresponding row. A number of different implementations has been shown in [56,58,74,75]. …”

Skeleton-supported stochastic networks of organic memristive devices: Adaptations and learning

“…Recently, learning in the memristor-based networks is under a wide discussion. [25][26][27][28][29][30][31] We assumed that the training must result in the reinforcement of the conductivity in one branch (S-D1) and to inhibit it in the other one (S-D2)…”

Transforming the short-term sensing stimuli to long-term e-skin memory

“…Among them, a metal/oxide/metal structure is one of the most studied device configuration for non-volatile memory applications with high endurance, long retention, fast operating speed, and potential compatibility to the complementary metal oxide semiconductor (CMOS) technology [9,10]. Memristive devices have also been studied for artificial neural systems to mimic the behaviour of synapses [11]. Despite significant relevance, the conventional integration strategy via CMOS technology is not valid for e-skin application and other strategy has seldom been explored.…”