When
Where
Speaker: Sarah Luca, Applied Mathematics GIDP
Title: Examples and Challenges of Spiking Algorithm Development for Neuromorphic Computing
Abstract: With the explosive rise in resource demands from advancements in AI/machine learning and large-scale scientific and edge computing applications, there is a growing need for more sustainable approaches to computing than conventional methods. Neuromorphic computing is an emerging field inspired by the brain which can perform highly parallel and energy efficient computations, making it an appealing alternative to conventional computing, but due to its emerging nature and differences in hardware design, neuromorphic algorithms are challenging to develop. In this talk, I will present examples of algorithms I developed for neuromorphic implementation on both digital and analog hardware that demonstrate the most common approaches to neuromorphic algorithm design. Throughout the talk, I will also highlight the many challenges I have observed while developing and implementing the algorithms in hardware and conclude with some future directions inspired by these results and observations.