Author:
Vishwa R,Karthikeyan R,Rohith R,Sabaresh A
Abstract
Abstract
Neuromorphic computing is a budding avenue though it has been known since the 80’s. The extensive research and development in the field of artificial intelligence particularly in the last decade is tremendous. The growth of artificial intelligence is expected to grow exponential in the years to come. Technologies like machine learning and IoT has made possible for many fields from industrial automation to business model prediction very affordable and far less complex. With growing digital devices, the number of devices connected to the cloud and in a network is doubling and in some cases are tripling in some ventures. Technologies like drones, autonomous cars, smart healthcare, smart cities and many other are moving towards more and more data and connected devices to the cloud. The present hardware system is at the verge of giving away as the data generation rate and processing volumes of the same is becoming a challenge. The hardware of today, though are advance are simply not adequate to support the expansion rate of growth of artificial intelligence in all fields. Increased devices result in increase data, increased processing raising challenges for current storage devices and processing devices. Neuromorphic chips, which promise to overcome this challenge, are currently being researched extensively by many computer giants who fear the future incompetency of hardware of which IBM is a major player. Ground breaking research in the field of memristor and artificial synapse have paved the way for neuromorphic chips which are expected to revolutionized the field for the better. This paper deals with the current research, physical and technical limitations and future scope of neuromorphic chips. The significance of memristor and artificial synapse towards neuromorphic computing is also dealt in detail.
Cited by
3 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献