To develop some kind of common sense for AI has always been the ultimate goal of AI research, ever since the first computers appeared about 60 years ago. Recently the increase in computing power reaching to sufficient levels, along with a greater number of new research projects on the subject are paving the way for such goals to be achieved.
One such project has recently started in University of Illinois at Chicago. The researchers were awarded a $500,000 project by National Science Foundation. The goal of the project is to create mathematical foundations and develop algorithms for building commonsense knowledge bases for AI that can evolve by itself. In other words, it is to enable a computer to learn about the world by supplying the necessary mathematical models for computation.
To process new incoming information for a computer in a reasonable way that will enable it to learn using common sense is an extremely difficult task. Even the word “reasonable” in this context need to be well defined at the first place, the researchers say. The successful completion of the project might enable robots, computers and other automated devices to be able to perfom tasks that are well beyond their current capabilities.