A research team having a lead IBM team and professors from the University of California have developed experimental computer chips that are capable of mimicking the abilities of human brain such as action, cognition and perception.
The successful outcome of this research could help in reducing power consumption and space occupied by processors used in the computers.
The professors participating in the research were Chris Kello and Stefano Carpin, who have been involved in leading the design and implementation of virtual environments for testing these chips. The fundamental building blocks of these chips include being able to learn through past experiences, create hypotheses, find correlations and learn various outcomes by mimicking the human brain’s synaptic plasticity and structural functioning. The research aims at developing a system capable of analysing complex information from multiple outputs received from sensors and also rewire itself during its interactions with the external environment. In order to achieve this outcome various techniques from nanoscience, computer science, cognitive science and neuroscience are merged together.
A cognitive scientist from the UC Merced’s School of Social Sciences, Humanities and Arts, pointed out that scientists have been working on cognitive theories and intelligent algorithms for many years but have failed to give proper importance to the role of the human brain. UC Merced has recently been awarded a financial grant for the second phase of the project, which is called Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNapse) from Defense Advanced Research Projects Agency (DARPA). The first two prototype chips are ready and are currently being tested. The project leader for IBM Reseach, Dharmendra Modha, explained that the research findings could be extended in future to designing traffic lights that are capable of integrating sights, sounds and smells and indicate unsafe intersections much ahead of the happening of disasters or for designing cognitive processors that could be used in servers, laptops and phones for improving their interaction with the environments.