The test, which was developed by Stanford University in the US, is one of the most authoritative ways to assess the machine's skills in understanding what has been read.
Chinese Internet company Alibaba Group has created a model of artificial intelligence, which for the first time managed to cope with the test of understanding the material better than man.
This was reported by Bloomberg.
It is noted that the test is one of the most authoritative ways to assess the machine's skills in understanding what has been read. It includes over 100,000 questions that were based on 500 articles from the online encyclopedia „Wikipedia.“ In addition, this test helps determine whether machines can process large amounts of information, as well as give accurate answers to the questions posed.
It turned out we are not that smart.
According to the research results, the artificial intelligence index is estimated at 82.4, while the mark in the human was at 82.3. The team at “Alibaba’’ noted that this is the first time that a machine managed to beat a human being in a similar test.
Artificial intelligence, developed by Microsoft, also managed to exceed human performance – its result in the same test was 82.7. However, these results were summed up a day later.
Artificial intelligence is considered to be complex systems that can not only act according to the program laid down by man, but also make effective self-learning, as well as actions that were not originally programmed by specialists.
Discussion about this post