Influence of the input signal's phase modulation on the performance of optical delay-based reservoir computing using...

Area of Science:

• Optics and Photonics
• Computational Science
• Machine Learning

Background:

• Semiconductor lasers with delayed feedback are effective for photonic reservoir computing.
• Input data is typically optically injected into the reservoir.
• The encoding method of optical injection signals influences system performance.

Purpose of the Study:

• To investigate the impact of different optical injection signal configurations on photonic reservoir computing performance.
• To compare the effectiveness of phase modulation versus amplitude modulation for data encoding.

Main Methods:

• Numerical simulations were employed to model the system.
• Different input configurations using Mach-Zehnder modulators and phase modulators were simulated.
• A one-step ahead time-series prediction task was used for evaluation.

Main Results:

• Performance is highly dependent on the information encoding in the optical injection signal.
• Modulating the phase of the injected signal resulted in significantly better performance compared to amplitude modulation.
• The study identified optimal input configurations for improved computational efficiency.

Conclusions:

• Phase modulation of optical injection signals is a superior method for enhancing photonic reservoir computing performance.
• Careful selection of input encoding strategies is crucial for optimizing complex problem-solving capabilities.
• This research provides insights for designing more efficient photonic reservoir computing systems.