This second edition of Neuromorphic Computing Principles and Organization offers a deep dive into the principles and organization of neuromorphic computing, emphasizing the design of fault-tolerant, scalable hardware for large and medium scale spiking neural networks with learning capabilities. Each chapter now includes exercises to enhance understanding and application. All existing chapters have been meticulously revised and improved, and a new chapter on advanced neuromorphic prosthesis design has been added, serving as a comprehensive case study for neuromorphic system applications.

The book starts with an overview of neuromorphic computing systems and delves into the fundamental concepts of artificial neural networks. It then discusses the evolution of artificial neurons in representing biological neuronal dynamics. Following this, it covers implementing these neural networks in neuron models, storage technologies, inter-neuron communication networks, learning mechanisms, and various design approaches. Fundamental design principles for building efficient neuromorphic systems in hardware are presented, along with a detailed discussion of the challenges in constructing spiking neural network architectures with numerous synapses. The book also explores learning in neuromorphic computing systems and the emerging memory technologies that hold significant promise for the field.

A dedicated chapter addresses the circuits and architectures used for communication in neuromorphic systems. This includes introducing the Network-on-Chip (NoC) fabric for receiving and transmitting spikes following the Address Event Representation (AER) protocol and memory access methods. It also covers interconnect design principles to help understand on-chip and off-chip communication concepts. Furthermore, the chapter introduces photonic interconnects for neuromorphic chips, along with advanced on-chip interconnect technologies such as silicon-photonic three-dimensional interconnects and fault-tolerant routing algorithms, which are discussed in detail.

The book also tackles reliability issues, describing several recovery methods for multicore neuromorphic systems essential for reliable processing in embedded neuromorphic applications. Subsequent chapters explain a reconfigurable design approach that supports multiple target applications through dynamic reconfigurability, network topology independence, and network expandability. The book concludes with a case study on the hardware-software design of a reliable three-dimensional digital neuromorphic processor geared explicitly towards the three-dimensional structure of biological brains. This platform enables high integration density and minimal spike delay of spiking networks and features a scalable design. Methods for fault detection and recovery in neuromorphic systems are also presented.

Neuromorphic Computing Principles and Organization is invaluable for researchers, scientists, advanced graduate students, and hardware-software engineers in neuromorphic chip hardware design. It is also an excellent teaching tool for imparting fundamental concepts, organizational strategies, and practical hardware software design techniques for reliable neuromorphic systems with learning and fault-tolerance capabilities.

Book Information
ISBN 9783031830884
Author Abderazek Ben Abdallah
Format Hardback
Page Count 320
Imprint Springer International Publishing AG
Publisher Springer International Publishing AG