Longing for Questions

Gabriel S.

The International Business Machines Corporation (IBM) developed a computer: Watson. Watson is an artificially intelligent computer system capable of answering questions. The difference between a document search and QA (question answering) is that document searching like Google: it takes keywords and finds them on webpages containing the keywords. The webpage is also ranked on its relevance and the source. Webpages do not provide the answer rather provide information to find the answer. The job for humans is to gather the information and provide an answer. Watson uses connections to take the job of humans. Could this mean a future with robots that can think?

127043797_1280x960

Watson’s Logo

Appearance in Jeopardy!

Watson was put to the test: thinking like a human. February of 2011, the Jeopardy! games began. Watson competed with Ken Jennings and Brad Rutter, two of the most successful players in the game show. The final score of two rounds were Watson with $77,147, Ken with $24,000 and Rutter come short with $21,600. IBM donated all winnings to respective charities along with 50% of Ken and Brad’s. Two members of the United States Congress challenged Watson, and Watson surprisingly lost the second round, though it still remained first for the Final Jeopardy clue. A disadvantage in the game is that Watson cannot detect other answers given by the other contestants. Thus, Watson giving answers twice or giving the right idea in the wrong “way.” For instance, the correct answer might have been: “What is: he’s missing a leg?” Watson gave the answer “What is leg?” Because Watson is not human, it could not responded to the mistake of another contestant.

Ken Jennings, Brad Rutter

Watson Answering in Jeopardy!

Watson’s Answer

When the clue was given, Watson analyzed the clue. Although computers detects codes as ones and zeros, the clue is given with nouns, verbs, adjectives etc. Watson parts the clue into parts of speech. This, helping it identify the relationship between each word. This gives Watson two things: what kind of question it is and what the question is asking for. Watson then searches through millions of documents to find as many answers as possible. It narrows the answers by seeing if each answer is supported with good evidence or bad evidence. Each document is analyzed by the relevance of it,the date and location. Each evidence helps provide rankings. If the ranking is high, Watson is more confident in answering a Jeopardy question. If the ranking is less than 50%, than Watson will not answer. This happens within seconds.

Future Applications

IBM is planning to use Watson for medical and finnicial. Watson is now deciding the fate of patients in hospital. Recieveing data, facts and previous experiments can help make the best decisions for patients. Watson can now save lives. Collecting data from previous houseowners can help make wise finnicail decisions. Although Watson may only answer questions, artificial intelegence can lead to much more than an answer. It can start the application of AI to robotics. Making more humanlike machines was never possible but now, it is just years away.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s