Single photonic reservoir dual-task simultaneous computing based on a semiconductor ring laser with filtered optical...

Area of Science:

• Photonics
• Optical Engineering
• Computational Science

Background:

• Reservoir computing (RC) offers a powerful framework for processing complex time-series data.
• Semiconductor ring lasers are promising candidates for photonic RC due to their rich dynamics.
• Enhancing the stability and robustness of photonic RC systems remains a key challenge.

Purpose of the Study:

• To propose and investigate a single photonic reservoir based on a semiconductor ring laser with filtered optical feedback (FOF).
• To achieve simultaneous reservoir computing using multiplexed directional modes.
• To enhance the robustness of reservoir computing operations.

Main Methods:

• Utilizing a semiconductor ring laser with filtered optical feedback (FOF) incorporating a Fabry-Perot filter.
• Performing numerical investigations of laser dynamics under FOF conditions.
• Evaluating reservoir computing performance on benchmark tasks: complex time series prediction and pattern classification.

Main Results:

• Filtered optical feedback (FOF) establishes a dynamically stable, steady state for both directional modes simultaneously.
• FOF expands the parameter ranges for preferred reservoir computing (RC) states, especially with optimized detuning.
• RC operations demonstrate significantly enhanced robustness against parameter fluctuations in both single-task and dual-task computing.

Conclusions:

• The proposed FOF system enables simultaneous dual-mode reservoir computing with improved stability.
• Filtered optical feedback enhances the robustness of photonic reservoir computing against parameter variations.
• This approach offers a promising pathway for developing more reliable and high-performance photonic computing systems.