This book explores the pivotal role of silicon-based digital processors in catalyzing the contemporary AI revolution. Despite the foundational groundwork for deep learning laid in the previous century, it was not until the utilization of graphics processing units that the training of expansive AI models became feasible. Nonetheless, the burgeoning demand for computing power poses a fresh hardware bottleneck. In this context, analog computing has resurfaced as a promising framework, serving dual purposes: (I) fostering the development of low-power neuromorphic accelerators for AI and (II) facilitating novel computing paradigms.

Divided into three parts, the book discusses what analog computing is and why it has the potential to be the key enabling technology for the next generation of AI hardware. The first part gives an overview of modern AI models, outlining their current challenges. The second and third parts describe the most recent advancements in electronic and photonic neuromorphic processors. This book aims to introduce AI and analog devices at various levels of analysis to anyone interested in the field, from science enthusiasts to graduate students and professionals. Its structured and pedagogical approach also makes it a prime candidate for textbook adoption in suitable university-level courses.

Lorenzo De Marinis has worked on photonic neuromorphic computing since the end of his MS. His main research activity concerns the conceptualization, design, and testing of novel photonic integrated circuits (PIC) with various integration technologies (SOI, InP, LNOI, SiN). With a background in electronic engineering, he also manages the electro-optic codesign of photonic systems and sub-systems. During his Ph.D., he developed a background in analog neuromorphic processors, machine learning, AI for optical communications, deployment of deep learning models in resource-constrained scenarios, and AI for in-networking applications. Currently, he is working as an assistant professor developing PICs for quantum random number generation, quantum computing and neuromorphic processing, under the Italian National Quantum Science and Technology Institute.