Neuromorphic Computing

Neuromorphic computing
Name of Student
Professor’s Name
Neuromorphic computing
Neuromorphic computing is an emerging field of technology which integrates human brain activities with electronic circuits, for providing viable outputs. In this mode of approach, huge VLSI systems (Very-Large-Scale-Integration) systems are integrated into the design. These systems implement electronic circuits (analogue circuits based on Boolean operations) to mimic the neural pathways present in the nervous system of human beings. Presently, research in neuromorphic computing is being carried out with analog, digital and mixed-mode systems. These systems are implemented to simulate and mimic behaviors related to perception, motor function and multi-sensory integration phenomenon (Monroe, 2014).
Oxide-based memristors, threshold switches, and transistors have been devised in the hardware systems to associate human anatomical model of the nervous system. The success of neuromorphic computing depends on the understanding and implementation of knowledge regarding the arrangement of neurons and their connections in the brain of a human. Such knowledge would be helpful to understand the functional characteristics of various neural pathways in sending and responding to stimulus ( Pickett, Medeiros-Ribeiro, &Williams, 2012).
It would include the cognitive capacities, the capacity of the brain to make complex calculations, the learning functions (acquiring new knowledge) and the process of responding to probable damage (neurodegeneration).

The discipline also incorporates the synaptic plasticity that is being created over a span of time. Further such models incorporate the evolutionary changes, which have taken place and may be anticipated. Thus, neuromorphic computers would be based on autonomic robotics, whose function would be to respond to situations without the need for human control. Moreover, they would be capable of taking individual decisions and would incorporate learning behaviors. The development of neuromorphic computing depends on the successful integration of biological sciences, physical sciences, information technology and electronic engineering. Therefore, such systems would be viewed as artificial systems with independent functioning capabilities (Waldrop, 2013).
Researchers all across the world are evaluating the prospects influenced by biological principles. A recent advancement has been made with the traditional Neumann architecture. The Neumann architecture is a logic core that operates in a sequential manner from the data received from the memory. This logic has been implemented in the neuromorphic computing process, where computation and memory are distributed over an array of simulated neurons through synaptic connections. The current projects are based on modifications of this architectural framework and programming is done to incorporate the functioning of the nervous system in the artificial intelligence system. A neuromorphic system called Neurogrid has been developed at Stanford University. It comprises of 16 –custom designed chips called Neuro cores. This neuro core has an analog circuit that contains neural elements for over thousands of neurons with computational capabilities, in an energy efficiency fashion. The emulated neurons have been connected through neural circuitry that is destined to utilize the spiking throughputs to the maximum potential (Benjamin et al., 2014).
Georgia Tech published a circuitry array in 2006. The chip was the first of its kind to establish the complex arrays of floating gate transistors. These transistors enabled the programmability of charge on the gates. Such circuitry has established in performing and mimicking the ion-channel ( ion-gated channels) of various neurons. These channels are required for establishing the resting membrane potential and depolarization (excitability of the neurons). Opening and closing of ion channels are a prerequisite for developing action potential in a nerve fibre. Researchers at MIT developed a circuit based on a chip that mimicked the synaptic transmission of neurons. In 2012, Spintronic Researches at Purdue established the role of memristors and lateral spin valves. They defended that such circuitry mimicked the human brain and functioning of neurons in the context of processing ability of the brain. Such arrangements were speculated to be more energy efficient than the conventional chips. The Human Brain Project is currently in the way of simulating total brain functioning. The project is using supercomputers along with huge amount of biological data (Maan et al., 2015).
Implications of Neuromorphic Computing
Neuromorphic Computing could be put to use to study the development of brain diseases. The changes in the brain and coping mechanism against such diseases could be implicated. The possible modifications in the synaptic zones and development of plasticity would be important to assess long-term changes in brain anatomy and its respective manifestations. .Such form of computing will not only be created to induce artificial intelligence but will serve as potential simulating models to understand neuropathology. Neuromorphic computing may be in its earliest phases, and there might be various hindrances to assess and inculcate an artificial brain (Zhao et al., 2010). However, advances in bioinformatics have already witnessed success with DNA logic gates. These logic gates have been created with Boolean operations based on ‘0” and “I” bits function. Hence, there is no reason to believe why neuronal models cannot be predicted and implemented. Moreover, neuromorphic computing will provide the necessary artificial intelligence, required in case of precision surgeries. The decision-making process could be inbuilt in the robotics system. Currently, robotic surgery is implemented and restricted to the current knowledge of anatomical structures within an individual. Creation of “neuromorphic surgeons” would not only operate on a patient but will ensure adequate troubleshooting and decision-making process, beyond clinical intelligence (Maan et al., 2015).
References
Benjamin, Ben Varkey; Peiran Gao; McQuinn, Emmett; Choudhary, Swadesh;
Chandrasekaran, Anand R.; Bussat, Jean-Marie; Alvarez-Icaza, Rodrigo; Arthur, John V.;
Merolla, Paul A.; Boahen, Kwabena (2014). “Neurogrid: A Mixed-Analog-Digital
Multichip System for Large-Scale Neural Simulations”. Proceedings of the IEEE 102(5),
699
Maan, A.K.; Kumar, D.S.; Sugathan, S.; James, A.P. (2015). “Memristive Threshold Logic
Circuit Design of Fast Moving Object Detection’ IEEE Transactions on Very Large Scale
Integration (VLSI) Systems 23 (10), 2337–2341.
Monroe, D. (2014). “Neuromorphic computing gets ready for the (really) big
time”. Communications of the ACM 57 (6), 13–15
Pickett, M. D.; Medeiros-Ribeiro, G.; Williams, R. S. (2012). “A scalable neuristor built with
Mott memristors”. Nature Materials 12 (2), 114–7
Waldrop, M. Mitchell (2013). “Neuroelectronics: Smart connections”. Nature 503 (7474),
22–4.
Zhao, W. S.; Agnus, G.; Derycke, V.; Filoramo, A.; Bourgoin, J. -P.; Gamrat, C. (2010).
“Nanotube devices based crossbar architecture: Toward neuromorphic
computing”. Nanotechnology 21 (17), 175202