Deep Cognitive Networks: Enhance Deep Learning by Modeling Human Cognitive Mechanism

Deep learning models and the human cognitive system process visual information differently, leading to a performance gap. To mimic this, deep cognitive networks model various cognitive mechanisms from cognitive neuroscience, such as attention, memory, reasoning, and decision, for more effective information processing. This book unifies these models and studies two representative cases of image-text matching.

Format: Paperback / softback
Length: 62 pages
Publication date: 31 March 2023
Publisher: Springer Verlag, Singapore

Despite the remarkable advancements made by deep learning models across a wide range of domains, including vision, speech, language, planning, control, and numerous others, a significant performance gap persists between these models and the human cognitive system. Researchers argue that one of the primary factors contributing to this gap is the distinct manner in which deep learning models and the human cognitive system process visual information. In an effort to bridge this performance gap, a trend has emerged since 2014 to model various cognitive mechanisms derived from cognitive neuroscience, such as attention, memory, reasoning, and decision, using deep learning models. This book aims to bring together these novel types of deep learning models and categorizes them as deep cognitive networks, which simulate various human cognitive processes based on deep learning models. Consequently, a wide range of cognitive functions, including selective extraction, knowledge reuse, and problem-solving, are implemented to enhance information processing efficiency.

This book begins by summarizing existing evidence of human cognitive mechanism modeling from cognitive psychology and presenting a comprehensive framework of deep cognitive networks that takes into account multiple cognitive mechanisms. Subsequently, it analyzes related works, with a particular focus on the taxonomy of four key cognitive mechanisms: attention, memory, reasoning, and decision, which are central to deep cognitive networks. Finally, the book studies two representative cases of applying deep cognitive networks to the task of image-text matching and discusses important future directions for this field.

In conclusion, this book provides a valuable contribution to the understanding and development of deep cognitive networks, which hold the potential to revolutionize the field of artificial intelligence by emulating human cognitive capabilities more effectively. By unifying these new models and exploring the taxonomy of key cognitive mechanisms, researchers can further advance our understanding of human cognition and develop more advanced intelligent systems that can solve complex problems and enhance our daily lives.

Weight: 131g
Dimension: 235 x 155 (mm)
ISBN-13: 9789819902781
Edition number: 1st ed. 2023