دانلود کتاب AI-Processor Electronics – فناوری پایه هوش مصنوعی

Computers have penetrated into every nook and corner of our homes, offices, industries, markets and entertainment media. They have been embraced and adored in all aspects of human life and activities from school homework assignments, online shopping and bank transactions to working from home. But computer progression and proliferation has now entered a new phase. From the epochs of mainframe to personal computing and the omnipresent laptops, tablets, smartphones and mobile devices, a new era of intelligent computers has dawned. These are the computers which can not only learn from experience, but think, reason and make decisions like intelligent human beings. Thanks to the artificial intelligence (AI) revolution, we are living in the age of intelligent machines that pass the Turing test in which an interrogator asks questions from a human and a machine but fails to ascertain which is a human and which is a machine. Pioneered in the 1950s, the dream of AI could not be realized at that time because of the non-availability of the required computing capabilities and the supporting database. AI evolved with the booming of data like a tsunami, maturation of AI algorithms and the leap forward in the semiconductor industry for microchip fabrication. Machine learning algorithms, particularly deep neural networks, pose unique challenges regarding the resources and execution time of training/inference calculations. The processing unit or processor of a computer is its principal component which is the point of convergence of all computational activities and on which the performance of the computer crucially depends. Therefore, it is the focus of attention of all computer scientists and engineers. All-purpose processors like the central processing unit (CPU) will always remain at the helm of affairs but they are deficient in dealing with the complex computations of the AI age. Therefore, development of specialized processing units for computationally-intensive AI algorithms is the hot topic of research worldwide. Particularly, without the stupendous progress in electronic processing units for computing, the great strides made by AI, machine learning and strikingly its subset ‘deep learning’ in recent years would not have been possible. Several off-the-shelf and emerging processors from different manufacturers have been used. These processors differ in their performance and energy efficiency. They are used for initial development and refinement of AI algorithms known as the training process as well as to apply trained AI algorithms to real-world data inputs to solve practical problems, viz., the inference phase. Electronics of AI is a subject whose impact on modern society needs hardly to be emphasized. It is out of the question to ignore the colossal impact of AI processor electronics on the advancement of AI because AI computing is not a bed of roses. It goes without saying that progress in AI largely depends on the innovations in AI processor electronics. Growth of AI applications will slow down with slackening of progress in AI processor technology. Being an area of research and development, books on this topic are few. Most of the information is only available on the websites of the manufacturers. So, a book on this topic is the need of the hour.