The molecular implementation a DNA based artificial neural networks
The field of Molecular Computing is an exciting area that aims to use biological molecules in order to create programmable gadgets. The idea was introduced in the 1990s, and several computer scientists and physical scientists have since realized it.
Researchers from East China Normal University (ECNU) and Shanghai Jiao Tong University (SJTU) have developed molecular convolutional networks (CNNs), based on synthetic regulatory DNA circuits. In a paper published by Nature Machine Intelligence they describe a new approach that overcomes many of the problems typically encountered in creating artificial neural networks using molecular components.
Hao Pei is one of the researchers involved in the study. She told TechXplore that the intersection of computer and molecular science is fertile ground for exciting new science. \”Especially the design of intelligent system is a goal of scientists for a long time,\” she said. The network size limits the computing power and scale of DNA neural networks compared to brains. Our primary goal was to increase the computing power of DNA molecular circuits using a neural network model.