Webinar Dr. Erika Covi 'Emerging devices for brain-inspired computing: A bridge between materials and circuits and systems'

In the past two decades, the shift towards a distributed computing paradigm led our smart systems to become more and more interconnected. These systems need to elaborate increasingly amount of data while featuring low-power operation, area efficiency, and ability to interact with the external world in real time. The answer on how to achieve this goal resides in our mind. Indeed, the human brain features a distributed system based on slow and unreliable components. Yet, it is able to learn from experience and compute unstructured data reliably with extreme energy efficiency. To achieve the performance of our brain’s real-time and low-power cognitive processes, we developed new computing paradigms, using mainly standard CMOS technology, such as neuromorphic computing and Spiking Neural Networks. In parallel, advanced learning theories enabling new ways of computation were formulated, and we investigated new technologies, such as memristive device technology, that employ new materials whose intrinsic properties emulate biological computational primitives. In particular, the development of non-volatile memristive devices provides an appealing power efficient approach and the emergence of volatile memristive devices offers a compact solution to reproduce temporal dynamics in the Neural Networks. These features, if correctly combined with the existing CMOS technology, can extend the functionality of current artificial intelligent systems. In this talk, we will discuss how memristive technology can enhance the performance of CMOS and what the challenges to build a memristive neuromorphic system are.