CMOS-Free Multilayer Perceptron Enabled by Four-Terminal MTJ Device
This addresses the need for more efficient and biomimetic neuromorphic hardware, though it appears incremental as it builds on existing spintronic device concepts.
The paper tackles the problem of implementing multilayer perceptrons in neuromorphic computing without CMOS circuitry by proposing a new spintronic neuron, enabling purely spintronic systems that simplify fabrication.
Neuromorphic computing promises revolutionary improvements over conventional systems for applications that process unstructured information. To fully realize this potential, neuromorphic systems should exploit the biomimetic behavior of emerging nanodevices. In particular, exceptional opportunities are provided by the non-volatility and analog capabilities of spintronic devices. While spintronic devices have previously been proposed that emulate neurons and synapses, complementary metal-oxide-semiconductor (CMOS) devices are required to implement multilayer spintronic perceptron crossbars. This work therefore proposes a new spintronic neuron that enables purely spintronic multilayer perceptrons, eliminating the need for CMOS circuitry and simplifying fabrication.