Explore bio-inspired neuromorphic circuit architectures in this 56-minute tutorial from tinyML EMEA 2021. Delve into the world of Neuromorphic Intelligence (NI) as Professor Giacomo Indiveri from ETH Zurich discusses how ultra-low power electronic circuits and brain-inspired computing architectures can overcome the limitations of current AI systems. Learn about strategies derived from neuroscience for robust, low-latency computation using electronic neural elements that mimic biological neurons and synapses. Discover applications of NI processing systems in extreme-edge use cases requiring low power and local data processing. The tutorial covers topics such as neural networks, memory and energy considerations, biological neural networks, silicon neurons, robust computation, and bioinspired architectures, providing insights into the future of computing and the potential of neuromorphic systems to revolutionize AI applications.
Bio-inspired Neuromorphic Circuits Architectures for Ultra-Low Power AI
Overview
Syllabus
Introduction
Neural networks
Memory and energy
Future of computing
Biological neural networks
Space and memory
Let time represent itself
Silicon neuron
Analog digital divide
Robust computation
Coefficient of variation
False myths
Bioinspired architectures
Dynams
spiking neural network
summary
closing remarks
Sponsors
Taught by
tinyML
