DNA Memristors and Their Application to Reservoir Computing.

Area of Science:

• Biomolecular Engineering
• Nanotechnology
• Computational Science

Background:

• Memristors are fundamental electronic components with memory properties.
• Existing memristor technologies face limitations in scalability and biocompatibility.
• Molecular and DNA-based systems offer potential for novel computing paradigms.

Purpose of the Study:

• To introduce novel molecular and DNA-based memristors.
• To demonstrate the application of these memristors in reservoir computing (RC) systems.
• To evaluate the performance of DNA-based RC systems for computational tasks.

Main Methods:

• Realization of molecular memristors using two molecular reactions with distinct rate constants.
• Implementation of DNA memristors using five DNA strand displacement (DSD) reactions.
• Development of a reservoir computing system utilizing a dynamical memristor array and molecular reactions for the readout layer.
• Application of bipolar fractional coding for state computation and readout mapping.

Main Results:

• Molecular and DNA memristors exhibit preserved hysteresis behavior, tunable via rate constant ratios.
• The state of the memristor is computable from output molecule concentrations.
• A DNA-based RC system successfully performed handwritten digit recognition and time series prediction.
• Performance of the DNA RC system was comparable to electronic memristor RC systems.

Conclusions:

• Molecular and DNA reactions can effectively realize memristive devices.
• DNA-based memristors are viable components for building sophisticated computational systems like reservoir computers.
• This work paves the way for bio-integrated computing and advanced molecular information processing.