Russia proposes to teach artificial intelligence to think differently
Scientists at Tyumen State University proposed the creation of a new model of artificial intelligence action based on the analysis of existing systems and a wide range of research in this area, which would insure against errors and enable artificial intelligence to think abstractly.
University scientists believe that as long as artificial intelligence does not have the ability to extend and evaluate an existing system in real time, there will be a risk that the decision made by artificial intelligence will be wrong.
To mitigate such risks, the university staff proposed an approach that would make artificial intelligence more "conscious." According to the authors, their model of artificial intelligence architecture combines the results of a wide range of physical, mathematical, cognitive, philosophical and other research in the field.
"We derived a key attribute of having a 'general artificial intelligence' system: it must be able to understand and be able to apply different models of its environment, that is, simply put, different 'worldviews' or 'theories'. Focusing on" theories ", as on certain data evaluation paradigms, the artificial intelligence will not only be able to" look at the situation differently ", but also to derive new principles, laws, appropriate for this area "- said a specialist at the School of Advanced Studies (SAS) University Louis Verwoort.
The proposed approach, the scientists are confident, will form the basis for the development of the next generation of artificial intelligence and will continue to work closely with experts in the field of machine learning to create specific computing solutions based on the new system