Computer Intelligence is an extension of the concept of Artificial Intelligence. It is the subject of William Meisel’s new book, Computer Intelligence: With Us or Against Us?
AI is sometimes defined as computers doing things that previously required humans. But when it is done for tasks such as optimizing energy uses of a bank of computer servers using millions of data items and machine learning, it is doing a task that no human could do.
Much of what is called AI today is machine learning using deep neural networks (DNNs), and one could argue that there is no conceptual breakthrough, just computer power passing a threshold where it is feasible to optimize such large data models and deliver them in a form that is sufficiently responsive that they can be used in practical applications.
One example is speech recognition. DNNs are used in today’s most powerful speech recognition, with models optimized from thousands of hours of speech data using many hours of processing. Once developed, the DNN with the optimized parameters can be executed with the optimal parameters sufficiently quickly to carry on a conversation without unacceptable latency.
The idea of “computer intelligence” is that the breakthrough that should be addressed is the passing of the power of computer technology beyond a threshold where using powerful technology such as DNNs is commercially practical. Computer power will continue to expand, and how Computer Intelligence (CI) will evolve should be examined in those terms, not as a challenge to human abilities. By looking at it that way, we will get insights not only into today, but tomorrow.
This web site presents some thoughts and news in the area of computer intelligence from William Meisel, PhD, who has worked in the area today called AI for most of his career and has been an industry analyst for more than 25 years, publishing newsletters and organizing conferences.