Forward-only learning in memristor arrays with month-scale stability

TL;DR

This paper demonstrates that memristor arrays can perform stable, energy-efficient on-chip learning using forward-only algorithms and reset-only updates, achieving near-backpropagation accuracy on ImageNet tasks with month-scale stability.

Contribution

It introduces a practical method for memristor array learning with forward-only algorithms and reset-only updates, enabling stable, low-energy, on-chip learning at array scale.

Findings

Achieved 89.5-89.6% accuracy on ImageNet-scale tasks.

Demonstrated stable analog states for at least one month.

Sub-1 V reset updates use 460x less energy than traditional programming.

Abstract

Turning memristor arrays from efficient inference engines into systems capable of on-chip learning has proved difficult. Weight updates have a high energy cost and cause device wear, analog states drift, and backpropagation requires a backward pass with reversed signal flow. Here we experimentally demonstrate learning on standard filamentary HfOx/Ti arrays that addresses these challenges with two design choices. First, we rely on forward-only training algorithms in the Forward-Forward family that use only inference-style operations. Second, we use sub-1 V reset-only, single-pulse updates that cut energy and yield stable analog states. We train two-layer classifiers on an ImageNet-resolution four-class task using arrays up to 8,064 devices. Two forward-only variants, two-pass supervised Forward-Forward and a single-pass competitive rule, achieve test accuracies of 89.5% and 89.6%, respectively; a reference experiment using backpropagation reaches 90.0%. Across five independent runs per method, these accuracies are indistinguishable within statistical uncertainty. Trained models retain accuracy for at least one month under ambient conditions, consistent with the stability of reset-only states. Sub-1 V reset updates use 460 times less energy than conventional program-and-verify programming and require just 46% more energy than inference-only operation. Together, these results establish forward-only, sub-1 V learning on standard filamentary stacks at array scale, outlining a practical, pulse-aware route to adaptive edge intelligence.