The evolution of Artificial Intelligence is a current topic, and it has actually been so for quite a long time. Professor Miquel Barceló, from UPC, explains that “informatics have always pursued the challenge of artificial intelligence. In the 1970s, icons were introduced to communicate with machines because there was no possible way of communicating in a more natural way”. In 1997, a computer called Deep Blue became historically famous when it won a chess game against the worldwide champion Garry Kasparov. Deep Blue used informatics to analyze in a few seconds the possible outcomes of every move, and chose the path to victory. Chess is a game with very strict moves, and therefore it is easy to understand for a computer. Human language is much more complex. When we talk, we often use implicit textual references and puns that are very difficult to understand for a computer. A clear example can be found in automatic translation, as the system replaces words in one language for their equivalents in another using a dictionary without analyzing the context in which they are used.
Watson is the new artificial intelligence project by IBM. The main goal is to understand questions given in natural language and reply correctly just like a human. This year, the team behind Watson has decided to test it in the US TV quiz show Jeopardy!, famous for its references, riddles and puns. In a special edition of the show that lasted 3 days, Watson competed against two famous contestants and won.
Further that competing in TV shows, the technology behind Watson can be used in areas like health to help people. Watson is capable of analyzing large bulks of data and extracting the important parts, and therefore could be useful to help doctors in diagnostics.