Utilizing rate-independent hysteresis for analog computing
Computational Physics
2025-03-17 v1
Abstract
Physical systems exhibiting hysteresis are increasingly being used in neuromorphic and in-memory computing research. Generally, the resistance switching of devices with rate-independent hysteresis are being investigated for their use as trainable weights in neural networks, whereas the dynamics of devices showing rate-dependent hysteresis are being investigate for their potential as nodes in, for example in reservoir computing systems. In our work we instead show the computing potential of a simple rate-independent hysteresis system. We show that by driving a system of only two linear branches with time-multiplexed inputs it is possible to generate nonlinear transforms and perform timeseries prediction tasks.
Cite
@article{arxiv.2503.11179,
title = {Utilizing rate-independent hysteresis for analog computing},
author = {Lina Jaurigue and Kathy Lüdge},
journal= {arXiv preprint arXiv:2503.11179},
year = {2025}
}