Neuromorphic Computing Principles and Organization

This book focuses on neuromorphic computing principles and organization, exploring how to build fault-tolerant scalable hardware for large and medium-scale spiking neural networks with learning capabilities. It covers artificial neural networks, neuron models, storage technologies, inter-neuron communication networks, learning, and design approaches. It also discusses the circuits and architectures used for communication in neuromorphic systems, including the Network-on-Chip fabric and advanced on-chip interconnect technologies. The book covers reliability and recovery methods for multicore neuromorphic systems and presents a reconfigurable design approach for multiple target applications.

Format: Hardback
Length: 244 pages
Publication date: 02 June 2022
Publisher: Springer Nature Switzerland AG

This comprehensive book delves into the realm of neuromorphic computing principles and organization, providing valuable insights into the development of fault-tolerant, scalable hardware for large and medium-scale spiking neural networks with learning capabilities. It offers a thorough exploration of the fundamental concepts of artificial neural networks, including artificial neurons and their representation of biological neuronal dynamics. The book then discusses the implementation of these neural networks in neuron models, storage technologies, inter-neuron communication networks, learning, and various design approaches.

The fundamental design principle to build an efficient neuromorphic system in hardware is explored, highlighting the challenges that need to be addressed in constructing a spiking neural network architecture with numerous synapses. Learning in neuromorphic computing systems and the major emerging memory technologies that promise neuromorphic computing are also covered in depth.

A dedicated chapter of the book focuses on the circuits and architectures employed for communication in neuromorphic systems. The Network-on-Chip fabric is introduced as a means of receiving and transmitting spikes following the Address Event Representation (AER) protocol, along with the memory accessing method. The interconnect design principle is also covered to provide a comprehensive understanding of on-chip and off-chip communication.

Advanced on-chip interconnect technologies, such as si-photonic three-dimensional interconnects and fault-tolerant interconnects, are discussed in detail to further enhance the performance and reliability of neuromorphic systems.

By presenting a comprehensive and in-depth exploration of neuromorphic computing principles and organization, this book serves as a valuable resource for researchers, engineers, and practitioners in the field of artificial intelligence, neuroscience, and computer engineering. It provides a solid foundation for developing innovative and efficient neuromorphic systems that can revolutionize the way we design and implement intelligent systems.

Weight: 559g
Dimension: 235 x 155 (mm)
ISBN-13: 9783030925246
Edition number: 1st ed. 2022