סמינר: ACRC Webinar
“Spiking Manifesto” – a new way to build and train artificial neural networks
Current AI models are a thousand times less energy-efficient than the brain. These models use artificial neural networks (ANNs) requiring GPUs for multiplication of huge matrices. In contrast, spiking neural networks (SNNs) within the human brain do not employ matrix multiplication and dissipate orders of magnitude less energy.
In this talk, Prof. Eugene Izhikevich presents the Spiking Manifesto: a framework for thinking about popular AI models in terms of spiking networks and polychronization, and for interpreting spiking activity within the brain as nature’s way of providing look-up tables. This approach offers a novel approach to develop ANNs in hardware and to build AI models based on innovative architectures. The spiking manifesto provides a path to thousandfold improvements in energy efficiency.
Bio:
Founder and CEO of SpikeCore, San Diego, California
Founder and Chairman of the Board of Brain Corp, San Diego, California
Founder and Editor-in-Chief of Scholarpedia – the peer-reviewed encyclopedia
Publications: https://scholar.google.com/citations?user=1JeBX-IAAAAJ
Important: Registration is required.
Zoom link will be sent after registration.