<p>Volatile mem-elements can operate at locally active steady states thereby internally amplifying energy fluctuations. Such elements can display persistent dynamical response to a constant excitation when coupled to an appropriate passive network. While such persistent oscillations have been demonstrated for memristors, similar work for memcapacitors and meminductors is currently missing. With both elements now physically realized, their realistic models may be developed and investigated for local activity, and coupling networks which lead to persistent dynamics can be designed. This work reports the fabrication of a volatile meminductor mathematically shown to operate at the “edge-of-chaos” and presents experimental evidence of passive coupling resulting in persistent oscillatory behavior. The meminductor demonstrates quantifiable “contours of inversion (COI)” originating from state-dependent inductance, thus highlighting that complexity cannot emerge without “mem”-behavior. Finally, the coupled meminductor system has been shown to emulate second order neuronal behaviors, thus experimentally affirming the potential of meminductors in neuromorphic computing applications.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Edge-of-chaos operation and persistent dynamics for neuromorphic meminductor computing

  • Abhiram Dinavahi,
  • H. Rusty Harris

摘要

Volatile mem-elements can operate at locally active steady states thereby internally amplifying energy fluctuations. Such elements can display persistent dynamical response to a constant excitation when coupled to an appropriate passive network. While such persistent oscillations have been demonstrated for memristors, similar work for memcapacitors and meminductors is currently missing. With both elements now physically realized, their realistic models may be developed and investigated for local activity, and coupling networks which lead to persistent dynamics can be designed. This work reports the fabrication of a volatile meminductor mathematically shown to operate at the “edge-of-chaos” and presents experimental evidence of passive coupling resulting in persistent oscillatory behavior. The meminductor demonstrates quantifiable “contours of inversion (COI)” originating from state-dependent inductance, thus highlighting that complexity cannot emerge without “mem”-behavior. Finally, the coupled meminductor system has been shown to emulate second order neuronal behaviors, thus experimentally affirming the potential of meminductors in neuromorphic computing applications.