Emulating dynamic synaptic plasticity over broad timescales with memristive device

Authors:

Abstract
Memristive devices have shown great potential in realizing artificial synapses efficiently for neuromorphic computing. However, emulation of various subtle forms of plasticity over broad timescales in a single device remains a challenge. In this paper, we designed a Ag/MgO/Pt memristive device exhibiting a volatile switching behavior and demonstrated various forms of synaptic plasticity from milliseconds to days in a single device. The volatile behaviors are owing to the formation and spontaneous rupture of silver nano-filaments with and without applying electrical stimuli. By manipulating the input voltage pulse strength, such as pulse amplitude, interval, and pulse number, several key features of biological synaptic plasticity including paired pulse facilitation, augmentation, post-tetanic potentiation, early long-term plasticity (LTP), and late-LTP were emulated. In addition, a reversible transition between short-term plasticity and LTP was also demonstrated. This work enables artificial synapses with rich plasticity in a single nano-device, providing a pathway to develop a large scale artificial neural network for brain inspired computing systems.
Funding Information
  • A*STAR | Science and Engineering Research Council Pulic Sector Research Funding (1521200085)
  • Temasek Laboratories @ Singapore University of Technology and Design Seed Research Programme (IGDSS1701051)

This publication has 34 references indexed in Scilit: