Neuromorphic Learning and Recognition With One-Transistor-One-Resistor Synapses and Bistable Metal Oxide RRAM

TLDR

Resistive switching memory (RRAM) has been proposed as an artificial synapse in neuromorphic circuits due to its tunable resistance, low‑power operation, and scalability. The study aims to validate bistable RRAM and present a compact one‑transistor/one‑resistor synaptic circuit using HfO₂ RRAM for high‑density neuromorphic systems. The authors implement this synapse and simulate a fully connected network that learns and recognizes patterns online with varying POST spike voltages. Spike‑timing‑dependent plasticity is demonstrated in both deterministic and stochastic RRAM regimes, and the simulations confirm that bistable RRAM enables high‑performance, online unsupervised pattern learning and recognition.

Abstract

Resistive switching memory (RRAM) has been proposed as an artificial synapse in neuromorphic circuits due to its tunable resistance, low power operation, and scalability. For the development of high-density neuromorphic circuits, it is essential to validate the state-of-the-art bistable RRAM and to introduce small-area building blocks serving as artificial synapses. This paper introduces a new synaptic circuit consisting of a one-transistor/one-resistor structure, where the resistive element is a HfO2 RRAM with bipolar switching. The spike-timing-dependent plasticity is demonstrated in both the deterministic and stochastic regimes of the RRAM. Finally, a fully connected neuromorphic network is simulated showing online unsupervised pattern learning and recognition for various voltages of the POST spike. The results support bistable RRAM for high-performance artificial synapses in neuromorphic circuits.