The computing carrier by electrons strictly restricts the computing capacity. However, intelligent electronic neural network hardware is still suffering from limited electrical bandwidth and huge energy consumption for the larger matrix decomposition. Nowadays, electronic neural network accelerators and processors based on the graphics processing unit (GPU), application-specific integrated circuits (ASIC), and field-programmable gate array (FPGA) dominate the commercial AI technique and specific function processing. Progress of intelligent hardware plays a crucial role in developing next-generation advanced neural network processors. With the rapid development of AI and increasing demand for high-capacity datasets processing, high-performance processors with accelerated matrix multiplication operations and high parallelism have attracted great attention in recent years. The DNNs based on commercial electrical hardware processors or specifically optimized algorithms are extensively explored in pattern recognition, intelligent translation system, and material science. This work provides a high-performance architecture for future parallel high-capacity optical analog computing.ĭeep neural network (DNN) has been an essential tool for developing general-purpose artificial intelligence (AI). Efficient parallelization feasibility with wavelength division multiplexing is demonstrated in our high-dimensional ONN. We successfully achieve the digit classification of different frequency components by demultiplexing the output signal and testing power distribution. Two frequency components from the microcomb source carrying digit datasets are simultaneously imposed and intelligently recognized through the ONN. Here, we demonstrate the dual-layer ONN with Mach–Zehnder interferometer (MZI) network and nonlinear layer, while the nonlinear activation function is achieved by optical-electronic signal conversion. Optical neural network (ONN) has the native advantages of high parallelization, large bandwidth, and low power consumption to meet the demand of big data. Parallel multi-thread processing in advanced intelligent processors is the core to realize high-speed and high-capacity signal processing systems.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |