Semi-Trained Memristive Crossbar Computing Engine with In-Situ Learning Accelerator

TL;DR

This paper introduces a semi-trained memristive crossbar architecture with in-situ learning for on-device neural network training, optimizing resource use and power efficiency, verified through system-level simulations.

Contribution

It proposes a semi-trained crossbar approach with partial memristor programming, enhancing resource efficiency and power savings for neural network hardware.

Findings

Power consumption is approximately 42.16μW for a 4x4 network.

Area footprint is estimated at 26.48μm x 22.35μm.

Applicable to various neural network architectures.

Abstract

On-device intelligence is gaining significant attention recently as it offers local data processing and low power consumption. In this research, an on-device training circuitry for threshold-current memristors integrated in a crossbar structure is proposed. Furthermore, alternate approaches of mapping the synaptic weights into fully-trained and semi-trained crossbars are investigated. In a semi-trained crossbar a confined subset of memristors are tuned and the remaining subset of memristors are not programmed. This translates to optimal resource utilization and power consumption, compared to a fully programmed crossbar. The semi-trained crossbar architecture is applicable to a broad class of neural networks. System level verification is performed with an extreme learning machine for binomial and multinomial classification. The total power for a single 4x4 layer network, when implemented in IBM 65nm node, is estimated to be ~ 42.16uW and the area is estimated to be 26.48um x 22.35um.