An integrated physical diffractive optical neural network (DONN) is proposed based on a standard silicon-on-insulator (SOI) substrate. This DONN has compact structure and can realize the function of machine learning with whole-passive fully-optical manners. The DONN structure is designed by the spatial domain electromagnetic propagation model, and the approximate process of the neuron value mapping is optimized well to guarantee the consistence between the pre-trained neuron value and the SOI integration implementation. This model can better ensure the manufacturability and the scale of the on-chip neural network, which can be used to guide the design and manufacturing of the real chip. The performance of our DONN is numerically demonstrated on the prototypical machine learning task of prediction of coronary heart disease from the UCI Heart Disease Dataset, and accuracy comparable to the state-of-the-art is achieved.

Machine learning technologies have been extensively applied in high-performance information-processing fields. However, the computation rate of existing hardware is severely circumscribed by conventional Von Neumann architecture. Photonic approaches have demonstrated extraordinary potential for executing deep learning processes that involve complex calculations. In this work, an on-chip diffractive optical neural network (DONN) based on a silicon-on-insulator platform is proposed to perform machine learning tasks with high integration and low power consumption characteristics. To validate the proposed DONN, we fabricated 1-hidden-layer and 3-hidden-layer on-chip DONNs with footprints of 0.15 mm2 and 0.3 mm2 and experimentally verified their performance on the classification task of the Iris plants dataset, yielding accuracies of 86.7% and 90%, respectively. Furthermore, a 3-hidden-layer on-chip DONN is fabricated to classify the Modified National Institute of Standards and Technology handwritten digit images. The proposed passive on-chip DONN provides a potential solution for accelerating future artificial intelligence hardware with enhanced performance.

No abstract

In this paper, a new architecture of optoelectronic convolutional neural networks (CNNs) based on time-stretch method is proposed. In this loop-shaped structure mainly composed of fiber optical and electronic devices, computations of data from each layer of CNN which are carried by light pulses with high repetition rate can be accomplished in a serial way. Therefore, a 5-layer CNN with two convolution layers, two mean pooling layers and one fully-connected layer are implemented. Under the test of handwriting digit recognition, its accuracy can reach up to 95% under ideal circumstances. Tests under different relative noise levels have been conducted and analyzed as well.

No abstract

No abstract

Machine learning technologies have been extensively applied in high-performance information-processing fields. However, the computation rate of current hardware is severely circumscribed by the conventional Von Neumann architecture. Photonic approaches have demonstrated extraordinary potential for executing deep learning involving complex calculations. In this work, an on-chip diffractive optical neural network (DONN) based on a silicon-on-insulator (SOI) platform is proposed to perform machine learning tasks with high integration and low power consumption. To validate the proposed DONN, we fabricated 1-hidden-layer and 3-hidden-layer on-chip DONNs with footprints of 0.15 mm2 and 0.3 mm2 and experimentally verified their performance in a classification task on the Iris plants dataset, yielding accuracies of 86.7% and 90%, respectively. The proposed fully passive on-chip DONN provides a potential solution for accelerating future artificial intelligence hardware with enhanced performance.

In this paper, an artificial-intelligence-based fiber communication receiver model is put forward. With the multi-head attention mechanism it contains, this model can extract crucial patterns and map the transmitted signals into the bit stream. Once appropriately trained, it can obtain the ability to restore the information from the signals whose transmission distances range from 0 to 100 km, signal-to-noise ratios range from 0 to 20 dB, modulation formats range from OOK to PAM4, and symbol rates range from 10 to 40 GBaud. The validity of the model is numerically demonstrated via MATLAB and Pytorch scenarios and compared with traditional communication receivers.