- Target Will Match Amazon Prices on the Spot!Here is How to Get that Amazon Price at Target - 8/8/2018
- Washington Post Columnist Suzi Parker Erroneously Reports Satirical Story About Sarah Palin Taking Host Position with Al Jazeera - 12/10/2016
- Android People App on Google 4.2 Operating System Jelly Bean Omits Entire Month of December from Calendar - 12/10/2016
Watson, an IBM computer (or supercomputer) that serves as a highly-advanced Question Answering system (QA system), is tied with Brad Rutter, the all-time Jeopardy! money winner (3.2 million) who has never lost a game of Jeopardy!, after the first round of the game show concluded tonight.
Watson, who is named after IBM’s founder, Thomas J. Watson, and Rutter have 5,000 points each. Ken Jennings, winner of the most consecutive Jeopardy! games in history (74), is currently in third place with 2,000 points.
The three-day game has taken on huge symbolic value, with some billing tonight’s show as a type of “Man vs. Machine” battle. While this is of course an inaccurately dramatic way to look at the Jeopardy! showdown, legitimate questions are being asked about the ever-increasing role of computers in the future, and how this development will impact fundamental institutions of society.
Computers have long been known as heroes of raw computation, but to play Jeopardy! is not to crunch numbers. In addition to deep learning in many subjects, Jeopardy! contestants must also interpret word play and disregard irrelevant information in the questions that may be present for entertainment purposes only.
In tonight’s show, for example, there was a clue category titled “Alternate Meanings.” In this category, clues had two parts, each of which pointed to the same single-word answer. On Monday’s show, one part of a clue from the “Alternate Meanings” category was something like “sophisticated style,” and the other part was “a group of students who graduate at the same time.” The correct answer – “What is class?” – required contestants to recognize the two different senses of the word “class” hinted at, and before this, contestants had to make sense of the category as a whole, determining what correct answers must look like.
The Internet Patrol is completely free, and reader-supported. Your tips via CashApp, Venmo, or Paypal are appreciated! Receipts will come from ISIPP.
CashApp us
Venmo us
Paypal us
So, Watson cannot merely cull the internet for answers to questions, even if the machine were connected to the internet, which it’s not. (Search doesn’t work like that: humans have to not only come up with appropriate search terms to find an answer, but they also have to read documents and connect information. Moreover, all the information one is looking for is often not available in a single source.)
Since Watson is not even connected to the internet (or any other external source, for that matter), all of Watson’s data is self-contained. The data that Watson has access to is primarily natural-language text. In a very real way, you can think of Watson as having read and attempted to understand innumerable books and other text-based media, like an erudite human. This “education” is also how Watson developed a sense of word play and colloquialisms.
Watson was also able to prepare by playing around 100 practice games against former Jeopardy! players before appearing on tonight’s show.
Since Watson cannot understand information that is fed to it verbally or visually (Watson neither hears nor sees), the machine received clues in the form of written text, which was supplied concurrently as Alex Trebek read the clues to the human contestants.
Yet another complicating factor Watson must contend with is the fact that wrong answers are penalized on Jeopardy!, so Watson first has to search for information that may supply the correct answer to a devilishly clever clue, and then it must assign a probability to the pieces of information (Watson’s hypotheses, if you will) it retrieves. Watson will not buzz in to supply an answer unless it meets the confidence threshold, and the confidence threshold varies from clue to clue.
As an anonymous female voice put it at the start of the show: “Watson knows what it knows, and knows what it doesn’t know.”
This isn’t the first time IBM has taken on humanity. In 1997, an IBM machine by the name of “Deep Blue” beat Garry Kasparov, a chess grandmaster and former World Chess Champion who many consider to be the greatest chess player of all time. (Kasparov, for what it’s worth, remains deeply skeptical that Deep Blue’s play was not aided by human guidance during part of the match, which would have violated the rules of the game.)
Watson, not surprisingly, is an impressive machine. (Techies be warned: it’s about to get a little hot in here.) IBM describes Watson thus:
“Watson is an application of advanced natural language processing, information retrieval, knowledge representation and reasoning, and machine learning technologies to the field of open domain question answering. At its core, Watson is built on IBM’s DeepQA technology for hypothesis generation, massive evidence gathering, analysis, and scoring. Watson is a workload optimized system designed for complex analytics, made possible by integrating massively parallel POWER7 processors and the IBM DeepQA software to answer Jeopardy! questions in under three seconds. Watson is made up of a cluster of ninety IBM Power 750 servers (plus additional I/O, network and cluster controller nodes in 10 racks) with a total of 2880 POWER7 processor cores and 16 Terabytes of RAM. Each Power 750 server uses a 3.5 GHz POWER7 eight core processor, with four threads per core.”
Alex Trebek attempted to simplify this description, claiming at the start of the show that Watson’s computation power is roughly equivalent to 2,800 standard computers.
The Internet Patrol is completely free, and reader-supported. Your tips via CashApp, Venmo, or Paypal are appreciated! Receipts will come from ISIPP.
CashApp us
Venmo us
Paypal us
