Ferroelectric Devices as Physical Reservoirs: Enabling Nonlinear Dynamics and Memory in Neuromorphic Systems

Key Points

• The research aims to explore the use of ferroelectric devices as physical reservoirs for enhancing computational efficiency in neuromorphic systems.
• Utilized reservoir computing for system evaluation
• Examined intrinsic dynamics of ferroelectric devices
• Compared performance with recurrent neural networks
• Demonstrated effective memory retention in ferroelectric devices
• Showed improved training efficiency versus traditional recurrent neural networks
• Reported successful integration of nonlinear dynamics into computational tasks