研究目的
Investigating the high-speed neuromorphic reservoir computing system based on a semiconductor nanolaser with optical feedback under electrical modulation for the first time and demonstrating it numerically.
研究成果
The proposed high-speed neuromorphic reservoir computing system based on a semiconductor nanolaser under electrical modulation achieves an information processing rate of 10Gpbs. Larger Purcell factor F and spontaneous emission coupling factor β extend the range of good prediction performance. The system is robust for the feedback phase, providing theoretical guidelines for the design of RC-based integrated neuromorphic photonic systems.
研究不足
The study is a numerical demonstration, and practical implementation may face challenges such as controlling the feedback phase in real-life equipment.
1：Experimental Design and Method Selection:
The study employs a numerical demonstration of a high-speed neuromorphic reservoir computing system based on a semiconductor nanolaser with optical feedback under electrical modulation. The Santa-Fe chaotic time series prediction task is used to quantify the prediction performance.
2：Sample Selection and Data Sources:
The first 4000 points in the Santa-Fe data set are used, with the first 3000 points for training and the next 1000 for testing.
3：List of Experimental Equipment and Materials:
A semiconductor nanolaser with optical feedback under electrical modulation is used as the reservoir layer of the RC system.
4：Experimental Procedures and Operational Workflow:
The input states are sampled and held for a duration time T, equal to the time delay τ in the feedback loop. The input state is multiplied with a mask, resulting in a temporal input stream J(t).
5：Data Analysis Methods:
The prediction performance is quantitatively evaluated by adopting the normalized mean square error (NMSE) between the target value and the predicted value.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
