Riding the Wave of Green Change: Advancing Wave-Based Analog Reservoir Computing for Sustainable AI

Abstract

Wave-based reservoir computing stands as a promising analogue computing paradigm, offering the potential for efficient and sustainable AI hardware. Various systems, including hydrodynamic, random, nonlinear, quantum, and multimodal approaches, have demonstrated robust physical computing capabilities, each exhibiting reliable input-output characteristics within complex operational domains. In this presentation, I will explore the recent advancements in photonic and hydrodynamic reservoir computing and share the journey that led me to investigate the complexities of wave propagation, with the goal of understanding its pivotal role in facilitating machine learning. My objective is to convey three key takeaways: first, the imperative to align technological innovation with ecological responsibility to address climate challenges; second, the specific contributions I have made toward achieving this alignment; and third, the strategies I have employed to understand, measure, optimise, and control wave-based reservoir computing systems, with the overarching goal of fostering a brighter, greener future.

Our Speaker

Dr Giulia Marcucci is a deep-tech entrepreneur and physicist who has been working on sustainable AI long before it became urgent. She holds a PhD in Physics, specialising in photonics for new computing models. She has held research roles at the Universities of Rome, Ottawa, and Glasgow, and served as Algorithms Lead at a neuromorphic startup in London. A recognised expert in AI-photonic computing, she now leads LumiAIres as Co-Founder and CEO, bringing purpose-built, energy-efficient AI-photonic processors to market through co-designed AI software and optical hardware.