Where Is Watson Now Watson can do much more than compete on a quiz Imagine walking up to a computer telling it your show Now the same AI technology that played symptoms and then finding out what's wrong and on Jeopardy helps doctors treat patients New what the best treatment is all in natural language medical information comes out every day and it is IBM's Watson technology also helped write a impossible for doctors to read and learn about every cookbook Cognitive Cooking with Chef Watson new discovery They need an easy way to search IBM and the Institute of Culinary Education all that data The system clinicalKey uses Watson's gave Watson a database of online recipes DeepOA technology to search thousands of medical nutritional facts and flavor research Then journals textbooks and reports with this type of the supercomputer came up with entirely new technology a future with robot doctors is possible dishes including Italian pumpkin cheesecake The technology company IBM spent three years working on an AI technology called DeepQA that could answer natural language questions quickly and correctly The goal was to match or even beat human performance They chose Jeopardy as a test because the show uses natural language covers many different topics requires answers in less than three seconds and has penalties for wrong answers Watson used the DeepOA technology to answer questions during the quiz show appearance With each clue Watson used more than 100 different techniques to understand the clue decide how to search for answers and rank all the possible answers Each answer gets a confidence score that measures how likely it is to be correct To prepare for the show Watson trained hard working through more than 100 000 sample Jeopardy questions When it made a mistake the IBM team could teach Watson how to handle that type of clue
Doctors began using Watson to help determine treatment options for This Jeopardyl match took place in February 2011 over two days Jennings and Rutter cancer patients got lots of right answers but not enough to beat Watson The computer won easily with a final score of 77 147more than the two humans scores
Deep Blue Artificial Champions Garry Kasparov left plays chess against Deep Blue in 1997 Watson was neither the first nor the last Feng Hsiung Hsu one of Deep Blue's designers moves the computer program to defeat a human pieces for the computer opponent The checkers program Chinook was the first to win a human championship in 1994 Three years later an IBM program named Deep Blue defeated Garry Kasparov the world's chess champion In 2015 a group of computer scientists in Canada announced a computer program named Cepheus which plays an essentially perfect game of poker Chess is a game of perfect information meaning nothing is hidden from the players The rules are also very simple Each piece moves in one specific way A computer program with enough power could in theory test the results of every possible move and then choose the best one For example after each player has made one move there are 400 possible positions to calculate But that number quickly becomes enormous Twel moves ahead there are a quintillion 1 000 000 еее 800 000 possible game states most of those game states are ridiculous and a human player ignores them Deep Blue has been designed to ignore them too An algorithm measures the advantage of each possible move and the program follows only the paths that seem most promising when Deep Blue defeated chess champion Garry Kasparov the machine could search 200 million positions every second Now chess playing computer programs compete in their own tournaments and human chess players have learned new strategies from the machines Chinook Deep Blue Watson and Cepheus are each examples of AI Al shows up in many cutting edge technologies including robots driverless cars web searches and video games AI technologies use sophisticated algorithms or sets of instructions to solve a variety of difficult problems In science fiction Al characters are typically robots with superhuman brains In reality scientists still cannot build a robot that learns sees listens talks walks interacts and reasons as well as a person Instead there are many different technologies that 10
E e match A S O sease NSSSSSSSSSSSSSSSSSSSSSSS DEEP BLUE
