Impact of the leakage current of an AND-type synapse array on spiking neural networks. (October 2022)

• Record Type: Journal Article
• Authors: Jeon, Bosung; Song, Seunghwan; Hwang, Sungmin; Jang, Taejin; Park, Kyungchul; Park, Jonghyuk; Lee, Jong-Ho; Park, Byung-Gook
• Is Part Of: Solid-state electronics. Volume 196(2022)
• Publication Date: 2022-10
• Subjects: Neuromorphic -- AND-type synapse array -- Spiking neural networks; Semiconductors -- Periodicals; Semiconducteurs -- Périodiques; 621.38152
• DOI: 10.1016/j.sse.2022.108407
• Languages: English
• ISSNs: 0038-1101

Abstract

Highlights: Effect of the leakage current of an AND-type synapse array is investigated. Circuit-level modeling of a fabricated dual-gate synaptic device is conducted. As the input frequency lowers, the performance of a spiking neural network is severely degraded. Abstract: In this paper, we analyze the impact of the leakage current of an AND-type synapse array on a spiking neural network (SNN). In order to evaluate the firing rate modulation affected by the leakage current of a synapse array, the electrical characteristics of a fabricated dual-gate synapse are measured and modeled. A SPICE circuit simulation shows a severe change in the firing rate in the SNN, and a system-level simulation with MNIST datasets exhibits frequency-dependent performance degradation due to the leakage current. Compared to a case with negligible leakage current, a 6.98% accuracy loss at 50 kHz of input frequency, which is the firing rate of conventional CMOS neurons.