Constructing artificial neural networks using genetic circuits to realize neuromorphic computing

With the advent of the era of big data, existing computing systems limit the development of new technologies such as cloud computing and artificial intelligence. As the demand for high-performance computing continues to grow, traditional computing models are facing unprecedented challenges. Neural mimicry computing based on artificial neural networks provides a potential solution, and biological computing with advantages such as low energy consumption parallelization is very important for its research. in which gene circuitry will be the key to the construction of artificial neural networks. In this study, we used biological elements to construct gene circuits to form neural network structures to realize neural mimicry calculations. Based on the similarity between the structure of gene circuits and neural networks, we describe a neural network design based on the logic gate calculation of gene circuits. and realize the applications of linear classification. nonlinear classification and pattern classification. In the linear classification of NAND, the four datasets in the space arc divided into two categories by hyperplane. The structural design of the neural network for linear classification can be realized by using the genetic circuit, which includes live biological elements and the biological elements between different layers are connected by gene regulation. The simulation results of linear classification showed that the gene circuit realizes the function of NAND by means of the neural network and the output fluorescence value as the detection standard. In the nonlinear classification of XOR, the four damsels in the space arc divided into two categories. The structural design of the neural network for nonlinear classification can be realized by using the genetic circuit, which includes six biological elements. The simulation results of nonlinear classification showed that the gene circuit realizes the function of XOR by means of the neural network and the output fluorescence value as the detection standard. In the pattern classification. the neural network with five layers and eight neurons was constructed by a genetic circuit with eight biological elements connected by gene regulation. The simulation results of pattern classification showed that the gene circuit realizes the function of pattern classification by means of the neural network and the output fluorescence value as the detection standard. This study realizes neural mimicry computing by constructing neural network through engineering gene circuits. The integrated molecular computing system is expected to be applied to the manufacturing of artificial intelligence chips, and further applied to aerospace. information security and national defense construction, along with other fields.