Did anyone seriously believe that the humans could beat IBM's Watson computer in Jeopardy! last week? Did you seriously think that IBM would screwup the advertising coup of the year to Ken Jennings? I didn't even bother watching, but that does not mean I'm not impressed with the result. This was a major step since IBM's Chess contests with Gary Kasparov in 1996 and 97. Deep Blue beat Kasparov in May 1997, after first losing to him the previous year. That computers can play chess is no big deal these days, but understanding Jeopardy! clues, is another issue.
Next year (2012) is The Alan Turing Year and maybe IBM is planning to crack the Turing Test and win the Loebner Prize for the first computer to fool humans into thinking they are talking to a human, maybe. Passing The Turing Test would be a major step in Artificial Intelligence research and since it took 25 researchers and four years to teach Watson to play Jeopardy!, I have my doubts.
Both chess and Jeopardy! have much in common, they are both games, somewhat predictable with rules, although Jeopardy! is much broader in scope, and more difficult because of the language aspect, neither really compares well to real life situations. However if the "real-life" situation is somewhat confined to a particular occupation, Watson may become an interesting help mate. This link will open a TED webpage that features a discussion with one of Watson's creators and some of the implications of this type of technology.
It seems Canadian medical schools are being accused of bad-mouthing family practitioners at a time when a shortage of this type of physician is possible because of aging boomers like me. Medical schools seem to have a bias to specialties for their students instead of family practice, and as a result less than one-third of medical students show interest in that field.
Here is a possible role for Watson, that is, program it so that it might be able to interact with human patients in a medical practice. Could Watson in real time assist in diagnosing ailments while freeing up physicians to do physical exams or other things? One of the major complaints that I can see in Canadian medicine is face time with physicians so that they can hear the patient's whole story. Of course this is the value of the family practitioner, s/he can discover all aspects of patient health IF they had the time. Furthermore, computers never have bad days, spats with their spouse or partner, headaches or get sidetracked by chatty patients. Computers will execute their program, ask ALL the pertinent questions, and a nuanced computer like a Watson, could get a useful chunk of data that a human doctor could miss. It's a thought anyway, but not that far-fetched as you can see here.
Next year (2012) is The Alan Turing Year and maybe IBM is planning to crack the Turing Test and win the Loebner Prize for the first computer to fool humans into thinking they are talking to a human, maybe. Passing The Turing Test would be a major step in Artificial Intelligence research and since it took 25 researchers and four years to teach Watson to play Jeopardy!, I have my doubts.
Both chess and Jeopardy! have much in common, they are both games, somewhat predictable with rules, although Jeopardy! is much broader in scope, and more difficult because of the language aspect, neither really compares well to real life situations. However if the "real-life" situation is somewhat confined to a particular occupation, Watson may become an interesting help mate. This link will open a TED webpage that features a discussion with one of Watson's creators and some of the implications of this type of technology.
It seems Canadian medical schools are being accused of bad-mouthing family practitioners at a time when a shortage of this type of physician is possible because of aging boomers like me. Medical schools seem to have a bias to specialties for their students instead of family practice, and as a result less than one-third of medical students show interest in that field.
Here is a possible role for Watson, that is, program it so that it might be able to interact with human patients in a medical practice. Could Watson in real time assist in diagnosing ailments while freeing up physicians to do physical exams or other things? One of the major complaints that I can see in Canadian medicine is face time with physicians so that they can hear the patient's whole story. Of course this is the value of the family practitioner, s/he can discover all aspects of patient health IF they had the time. Furthermore, computers never have bad days, spats with their spouse or partner, headaches or get sidetracked by chatty patients. Computers will execute their program, ask ALL the pertinent questions, and a nuanced computer like a Watson, could get a useful chunk of data that a human doctor could miss. It's a thought anyway, but not that far-fetched as you can see here.