Let the creation of full artificial intelligence is still science fiction, sometimes computer algorithms outperform human in unexpected areas. Recently, the two neural networks developed by large corporations, were developed by Stanford University to test understanding of the text better than men.
The test includes over 100 thousand questions, based on 500 articles, and lets find out how good man or machine “learned” text. Until now, people were always in the lead, but this time the algorithms showed results 82,44 and 82,65 while people managed to score only 82,3 points. If it is not specified, it refers to the average value, which demonstrated people-competitors, or about the maximum, although contest it can be concluded that probably meant the second option.
In a relatively recent time, the neural networks were able to surpass humans in games like go and poker. Also, artificial intelligence was able to discover earth-like planet, until then, had escaped the attention of astrophysicists. And recently a group of researchers from Japan presented an algorithm that is able to “read minds” of people, creating images similar to those the participants mentally represent.
Sometimes the question of whether artificial intelligence is a potential threat to the well-being of mankind becomes the subject of the dispute is highly well-known people — in particular the threat it see entrepreneur Elon Musk and physicist Stephen Hawking. As shown by a recent survey, 38 percent of Russians agree with the statement that most available jobs in the foreseeable future will be with a high probability of replaced by robots.
The most interesting day in the “MK” in one evening newsletter: subscribe to our channel in the Telegram.