Reservoir Network With Structural Plasticity for Human Activity Recognition

2024-10-01

IEEE Transactions on Emerging Topics in Computational Intelligence

AI Summary

To address the demand for efficient time-series processing on edge devices, this work proposes a neuromorphic echo state network (ESN) chip supporting both structural and synaptic plasticity. Implemented in 65 nm CMOS, it is the first ESN hardware to integrate on-chip structural plasticity, enabling dynamic optimization of network topology and sparsity while preserving reservoir stability and enhancing continual learning capability. Unlike conventional fixed-topology ESN hardware, the proposed design achieves localized, adaptive, and ultra-low-power temporal modeling. Experimental evaluation demonstrates 95.95% average accuracy on human activity recognition and 85.24% accuracy in prosthetic finger control. The chip achieves a throughput of 6×10⁴ samples/second at only 47.7 mW power consumption, establishing a new trade-off between computational efficiency, adaptability, and energy efficiency for edge-based time-series inference.

Technology Category

• Machine Learning: Hardware-aware ML
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

The unprecedented dissemination of edge devices is accompanied by a growing demand for neuromorphic chips that can process time-series data natively without cloud support. Echo state network (ESN) is a class of recurrent neural networks that can be used to identify unique patterns in time-series data and predict future events. It is known for minimal computing resource requirements and fast training, owing to the use of linear optimization solely at the readout stage. In this work, a custom-design neuromorphic chip based on ESN targeting edge devices is proposed. The proposed system supports various learning mechanisms, including structural plasticity and synaptic plasticity, locally on-chip. This provides the network with an additional degree of freedom to continuously learn, adapt, and alter its structure and sparsity level, ensuring high performance and continuous stability. We demonstrate the performance of the proposed system as well as its robustness to noise against real-world time-series datasets while considering various topologies of data movement. An average accuracy of 95.95% and 85.24% are achieved on human activity recognition and prosthetic finger control, respectively. We also illustrate that the proposed system offers a throughput of 6×$m {10}^{oldsymbol{4}}$ samples/sec with a power consumption of 47.7 mW on a 65 nm process.

Problem

• Develop neuromorphic chip for edge devices
• Enable local learning with structural plasticity
• Achieve high accuracy in time-series tasks

Innovation

• Custom neuromorphic chip for edge devices
• Supports structural and synaptic plasticity
• Achieves high accuracy and low power