This book takes the reader on a journey from the biological brain to an unexpected frontier of physics: networks of light that learn by reshaping the very material they inhabit. Starting from the discovery of neurons and the dynamics of memory, it reveals the brain not as a machine, but as a living city—where connections grow, fade, and reorganize through experience. From there, the narrative moves into the world of photorefractive crystals, where beams of light carve their own paths, interact, and form self-organizing structures known as solitons. These structures become the building blocks of a new kind of artificial intelligence: solitonic neural networks, systems that do not simulate learning but embody it physically. Blending neuroscience, physics, and engineering, the book explores three key ideas: how the brain encodes information, why traditional AI struggles to replicate it, and how light itself can become a medium for memory and computation. Written with a clear and engaging style, it is accessible to readers without a technical background, while offering deep insights for students, researchers, and professionals. At its core, this is not just a book about technology—it is a reflection on how intelligence arises, and how closely it is tied to the physical world.

Alessandro Bile graduated in Theoretical Physics from Sapienza University of Rome, contributing to the VIRGO experiment for gravitational-wave detection, and in Composition from the "Santa Cecilia" Conservatory, with a thesis bridging gravitational-wave physics and music. In 2022, he earned his Ph.D. in Electromagnetism (Doctor Europaeus) at Sapienza, focusing on photonic artificial intelligence,  a line of research he continues to pursue in the mathematical modelling of neuromorphic optical systems and photorefractive solitons. A research fellow at Sapienza since 2019, he was a Visiting Researcher at the Femto-ST Institute (Université de Franche-Comté) under a BGF fellowship from the French government, and has presented his work, spanning metamaterials, AI for cultural heritage, and educational technology  at conferences across the World. Since 2022, he has taught Physics at Sapienza's Faculty of Medicine. He is the author, with Springer Nature, of two monographs on solitonic neural networks.