Self-organizing neuromorphic nanowire networks as stochastic dynamical systems

Area of Science:

• Neuroscience
• Materials Science
• Computer Science

Background:

• Neuromorphic computing seeks brain-like hardware efficiency.
• Memristive networks show promise for in-memory computing.
• Linking network dynamics to information processing remains a challenge.

Purpose of the Study:

• To model neuromorphic nanowire network dynamics.
• To understand the interplay of deterministic and stochastic effects.
• To quantify their role in physical reservoir computing.

Main Methods:

• Modeling neuromorphic nanowire network behavior.
• Applying the Ornstein-Uhlenbeck process framework.
• Analyzing stimuli-dependent trajectories and stochastic effects.

Main Results:

• Neuromorphic nanowire network dynamics are accurately modeled by the Ornstein-Uhlenbeck process.
• This model captures noise and jumps in network behavior.
• The framework allows quantification of deterministic and stochastic dynamics' contributions.

Conclusions:

• A unified modeling framework for neuromorphic networks is established.
• This approach facilitates understanding of information processing in memristive systems.
• It paves the way for novel computing paradigms leveraging both deterministic and stochastic dynamics.