Towards transferable lightweight neuromorphic computing through a model-free temporal-switch framework

2026-07-01

AI Summary

This work addresses the challenge of performance degradation in lightweight neuromorphic computing when deployed across diverse hardware platforms due to device-to-device variability. To overcome this, the authors propose a model-free Time-Switching (TS) framework that explicitly incorporates extensive device variability during training, enabling direct cross-device performance transfer without requiring retraining or calibration. Integrated with memristor-based reservoir computing, the proposed approach achieves strong results on both Mackey–Glass time-series prediction and spoken-digit classification tasks, attaining a classification accuracy of 92.4%. The framework’s generality and cross-platform applicability are further validated across multiple memristor types and configurations, demonstrating its robustness and practical potential for real-world neuromorphic deployment.

Abstract

Lightweight neuromorphic computing offers a promising route to efficient AI, with particular benefits for resource-constrained edge deployments. However, its scalable deployment that can reliably transfer the expected performance has long been hindered by device-to-device variations, which necessitate costly and repeated re-training on new copies and undermine the practical advantages. To address this issue, we introduce a model-free temporal-switch (TS) framework to improve the direct transfer performance, without post-training calibration or adjustment. The TS framework provides a methodology to incorporate a broader spectrum of devices in the training process. In the validation using memristor-based reservoir computing, it enables high performance on unseen devices with a directly transferred readout. It achieves improved prediction in the representative Mackey--Glass benchmark, and the accuracy of 92.4% in spoken digit classification. Its efficacy is validated across different memristor families and RC configurations. Theoretical analysis not only reveals the general computational mechanism underlying its efficacy, but also underlines its potential applicability to other physical platforms.

Problem

• neuromorphic computing
• device variation
• model transfer
• edge AI
• hardware deployment

Innovation

• temporal-switch framework
• model-free
• neuromorphic computing
• device variability
• transferable reservoir computing