Table of Contents

Abstract

Translate Article

The fi eld of artifi cial intelligence (AI) has grown tremendously in recent years with the advent of Deep Neural Networks (DNNs) that surpass humans in a variety of cognitive tasks. It is estimated that the digital universe is growing at a rate of approximately 60% per year. This abundance of data, coupled with exponential gains in compute power, has enabled remarkable computational tools that have improved our understanding of today's incredibly complex economies and societies. Moreover, it has ushered in a new era of computing: the cognitive, or AI, era, in which data is considered a new natural resource and AI algorithms are indispensable tools. This special issue of the IBM Journal of Research and Development builds a bridge from conventional distributed learning systems, to hardware solutions that minimize data movement from the compute to memory, to novel hardware approaches on the horizon for deep learning for both training and inference. We have a collection of ten contributions from leading authorities in their fi elds that summarize the state of the art and show original research results.