Nanowires Based Artificial Synaptic Networks for Building Neuromorphic Artificial Intelligence

Abstract

Artificial Intelligence is rapidly adapting to the change in magnitude of computing resources available to us. More powerful processing units are being developed everyday and the data-driven Artificial Intelligence is getting better along with them. However, conventional form of Artificial Intelligence still lags far behind in achieving its parent idea of mimicking human intelligence as the conventional architectures do not offer some necessary properties. This is exactly why the need of developing synthetic brain-like synaptic networks which could deliver them arises. Nanotechnology, at present, is the most viable resource at hand to build complex networks with billions of artificial neurons and highly interconnected synapses that are compact enough to be handled and meet energy constraints as well. Researchers have combined nanotechnology and neuromorphic computing to come up with nanowires based artificial synaptic networks that provide promising results in improving neuromorphic artificial intelligence to a great extent. It is perhaps the most important step in developing next generation artificial intelligence. With this paper, we discuss in-depth the nanowires, neuromorphic systems, synaptic plasticity, biological synapses, current computing technologies along with their shortcomings. We also discuss their need and use in building synthetic synaptic networks and try to interpret their need in building neuromorphic artificial intelligence. Furthermore, we discuss technical challenges and psychological barriers in relevant future research that we, as researchers, felt noteworthy.