Today, despite the popular level, no one knows for sure how an algorithm works. Artificial Intelligence or how it is created, the truth is that we are getting used to working with them practically daily thanks to their use in different applications such as Facebook self-labeling or the popular ones bots of Telegram that, as you surely know, allow us to insert GIFs in conversations.
Although the truth is that at the moment this new technology leaves much more to be desired than it can offer, the truth is that it advance and evolve by leaps and bounds. You have proof of what I say in the advances that are being developed by OpenAI where, as reported Quartz, have developed a new platform under the name of Universe through which an artificial intelligence system may be able to use a PC like any human being.
OpenAI, thanks to the Universe project, wants to make artificial intelligence much less dependent.
The idea, at least that is what emerges from the peculiar and complicated report, is that the artificial intelligence system has the ability to use a kind of virtual mouse and keyboard and even pay attention to what is displayed on the screen to interact with a PC in the same way that any of us can.
This new tool puts into practice a particular form of research on artificial intelligence known as'reinforcement learning'. In this way the algorithm can learn from its mistakes through billions of repetitions. Basically we want to achieve that, thanks to Universe, an artificial intelligence system is capable of, for example, search for information online o play a game online, in the same way that a human being can.
As he commented Ilya Skutsever, OpenAI Research Director:
The key skill we would like our artificial intelligence systems to have is general problem solving. This way you could take the algorithm and put it in a new world, where it would understand what is happening, find out what to do and then get it done.
Further information: Quartz