Conférence : « Photonique pour des technologies neuromorphiques efficaces et à grande vitesse »

Professeur qui invite : José Azana

Biographie :

Antonio Hurtado is Professor of Neuromorphic Photonics at the University of Strathclyde’s Institute of Photonics (IoP) and UKRI Turing AI Fellow. He created and leads Strathclyde’s Neuromorphic Photonics Lab, where he carries out research on photonics for high-speed and efficient brain-inspired technologies for information processing, sensing and AI.

His research is currently supported by programmes awarded by UK and EU funding agencies, including, among others, the EU Pathfinder Open project ‘SpikePro’ and the UKRI Turing AI Acceleration Fellowship Programme ‘PhotonAI’. He is also Strathclyde’s lead for the newly established UK Multidisciplinary Centre for Neuromorphic Computing.    

Résumé :

Photonic approaches emulating the powerful computational capabilities of the brain are receiving increasing interest for radically new paradigms in ultrafast neuromorphic (brain-like) information processing and artificial intelligence (AI). In this talk, I will review our research on neuromorphic photonic systems using artificial optical spiking neurons and spiking neural networks (SNNs) for computation.

I will discuss our experimental work with different photonic platforms, e.g., semiconductor lasers, resonant tunneling diodes, passive micro-resonators, and describe their potentials for high-speed and low-energy neuromorphic (spike-processing) photonic systems for use in strategic applications (e.g. image processing, data classification, time-series prediction). During the talk, I will also introduce our work on photonic Reservoir Computing (RC) systems and Extreme Learning Machines (ELMs), yielding excellent performance across complex computing tasks at ultrafast rates, and benefitting from hardware-friendly implementations.