Watson hit a minicomputer parallel computing. IBM's position on the mainframe is difficult to shake - the small Watson (founder of the old Watson's son) in charge of IBM period, IBM invested about 60 billion U.S. dollars, this NASA research and development funding over the year, and ultimately to Japan's NEC and France The Cray pushed out, laid the IBM mainframe market dominance.
14 years ago, IBM developed the computer continued challenges to the limits of human intelligence.
previously developed based on the deep blue IBM AIX operating system for use in business and government departments have made a lot of orders, IBM also hopes the system can be used in Watson's DeepQA medical services, consulting and other fields in.
In addition, IBM's global R & D team of Watson, a pattern also increased the possibility of winning.
IBM tried to bring the advantages of the mainframe to the minicomputer field. Watson is the flagship of the IBM minicomputer sales star 90 sets of parallel computing, machine will come up.
these can be achieved by the 90 units IBM servers, 360 and the IBM computer chip-driven R & D DeepQA system. IBM processor is configured for the Watson Power 750 family of processors, this is the current RISC (reduced instruction set computer) architecture, the most powerful processor - which eventually come to support Watson reliable answers in less than 3 seconds .
For humans, the rules are simple, but for Watson, it means that a number of challenges. Watson must first understand the natural language of the host, which is not available in dark blue; followed Watson need to analyze these languages, such as which is the irony, which is the pun, which is a conjunction, followed by questions to determine the meaning of the keyword Watson for related searches, and to assess the possibility of a variety of answers; the final choice of three possible answers to the most, when one of the more than 50% probability, the program starts, press the Responder Watson.
very detailed division of these groups, such as Haifa, Israel, the depth of the team responsible for the open domain Question Answering systems engineering the search process, Tokyo, Japan will be responsible for Watson in the Q & A connecting word meanings and words, IBM China Research Institute and Shanghai Branch is responsible for different resources to provide data to support Watson, as well as specialized research method and research strategy team game team, and other betting.
much like science fiction movies of the plot - plans to develop a robot hidden in the world, those who recruit uncanny talent. The difference is, IBM is clearly more practical significance.
all areas. Its rules are correct problems can be money, wrong answers will be upside down.
DeepQA these R & D talent development system ensures Watson can have a new mode of human-computer interaction, such as natural language can understand and analyze, in fact, the understanding of natural language has also been at the IBM research on the strengths of artificial intelligence.
Computer Watson (Watson) After three rounds of competition, eventually won the quiz show Watson to obtain 7.7147 million results, a huge bonus to three times the advantages of this program the most intelligent of the two human players thrown far behind.
fact, it is IBM's characteristics, it has been a successful model to follow the industrial age, such as more good at this manufacturing line mode, and graded by scoring the rewards and penalties to employees, so that It is more efficient in all teams.
Now, in the minicomputer market, IBM and HP's market share is roughly 2:1, while a minicomputer company SUN is now focused on low-end market.
IBM longer than the manufacturing of complex machines, there is such a smart addition to the server computer (minicomputer cluster), it looks just like the phrase can be a bit ironic that as the buzzwords - IBM, not a manufacturing philosophy company?
->
According to this definition, Watson is considered in artificial intelligence has taken on a step. obscure, then you can see what
machine.
major areas. diagnosis, then Watson will face is the risk of patients to court, which is an IBM in terms of the application being considered. the limits of human intelligence, but also to help the company go with Amazon, Google and Microsoft are competing to win the high ground of science and technology initiative in the future.
artificial intelligence in the 1960s, R & D stagnation for several years, scientists have found that if to simulate the human brain to the definition of artificial intelligence that would walk into a dead end.
No comments:
Post a Comment