Neuromorphic computing in the edge: merging cyber and physical

Abstract

In today’s emerging world, both humans and objects are continuously connected, collecting and communicating data. The rising number of applications relying on smart ICT technology includes autonomous vehicles, industry 5.0, biomedical wearables and implants, environmental sensing, smart houses and offices, etc. With all these data, local computation in the edge has become a necessity to limit data traffic and response latency. Embedding AI processing in the edge may add high levels of smart autonomy to these systems. Progress in nanoelectronic technology in combination with emerging neuromorphic, event-driven architectures with dynamic learning capabilities allow to do this in a power- and hardware-efficient way. This keynote will explore some solutions being developed today, and illustrate them with some practical examples of integrated circuits that are on the verge of merging the cyber and the physical worlds.