Metaplasticity in Multistate Memristor Synaptic Networks
This work addresses the challenge of enhancing continual learning and memory retention in neuromorphic computing systems, representing an incremental advance in domain-specific hardware.
The paper tackles the problem of improving information retention and classification in synaptic networks by using multistate metaplastic synapses based on memristors, achieving a 2.1x increase in the number of classifiable input patterns compared to binary synapses with at least 75% mean accuracy.
Recent studies have shown that metaplastic synapses can retain information longer than simple binary synapses and are beneficial for continual learning. In this paper, we explore the multistate metaplastic synapse characteristics in the context of high retention and reception of information. Inherent behavior of a memristor emulating the multistate synapse is employed to capture the metaplastic behavior. An integrated neural network study for learning and memory retention is performed by integrating the synapse in a $5\times3$ crossbar at the circuit level and $128\times128$ network at the architectural level. An on-device training circuitry ensures the dynamic learning in the network. In the $128\times128$ network, it is observed that the number of input patterns the multistate synapse can classify is $\simeq$ 2.1x that of a simple binary synapse model, at a mean accuracy of $\geq$ 75% .