Unsupervised local learning based on voltage-dependent synaptic plasticity for resistive and ferroelectric synapses

2025-10-28

AI Summary

To address the challenges of low-power, real-time adaptive AI deployment on edge devices, this work proposes an unsupervised online learning mechanism based on voltage-dependent synaptic plasticity (VDSP), the first to universally support three distinct memristive devices: filamentary TiO₂- and HfO₂-based synapses, and ferroelectric tunnel junctions (FTJs) based on HfZrO₄. Unlike conventional spike-timing-dependent plasticity (STDP), our approach eliminates reliance on precise timing-critical pulse circuits. Implemented within an in-memory computing architecture, it enables compact, low-overhead neuromorphic learning and incorporates a variability suppression strategy to enhance robustness. Evaluated on MNIST-related tasks, the method achieves >83% recognition accuracy using only 200 neurons—significantly outperforming existing memristor-based spiking neural networks. This demonstrates the feasibility and energy-efficiency advantages of heterogeneous memristor integration for edge intelligence applications.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Research challenges in human and human-AI computation

Abstract

The deployment of AI on edge computing devices faces significant challenges related to energy consumption and functionality. These devices could greatly benefit from brain-inspired learning mechanisms, allowing for real-time adaptation while using low-power. In-memory computing with nanoscale resistive memories may play a crucial role in enabling the execution of AI workloads on these edge devices. In this study, we introduce voltage-dependent synaptic plasticity (VDSP) as an efficient approach for unsupervised and local learning in memristive synapses based on Hebbian principles. This method enables online learning without requiring complex pulse-shaping circuits typically necessary for spike-timing-dependent plasticity (STDP). We show how VDSP can be advantageously adapted to three types of memristive devices (TiO$_2$, HfO$_2$-based metal-oxide filamentary synapses, and HfZrO$_4$-based ferroelectric tunnel junctions (FTJ)) with disctinctive switching characteristics. System-level simulations of spiking neural networks incorporating these devices were conducted to validate unsupervised learning on MNIST-based pattern recognition tasks, achieving state-of-the-art performance. The results demonstrated over 83% accuracy across all devices using 200 neurons. Additionally, we assessed the impact of device variability, such as switching thresholds and ratios between high and low resistance state levels, and proposed mitigation strategies to enhance robustness.

Problem

• Developing energy-efficient unsupervised learning for edge AI devices
• Implementing voltage-dependent synaptic plasticity in memristive synapses
• Addressing device variability in resistive and ferroelectric memory systems

Innovation

• Voltage-dependent synaptic plasticity for unsupervised local learning
• Adapted to three types of memristive devices with distinct switching
• System-level simulations validate MNIST recognition with high accuracy