I congratulate all the individuals involved in the years of research and development of algorithms, technologies, and systems that contributed to the very impressive culmination we saw when the IBM computer system, Watson, won on Jeopardy last week. For younger readers – Jeopardy is a week-night quiz show that has been on TV since 1964, where three very, very bright human contestants select topical categories and then buzz in to provide a ‘question’ to the brief ‘answer’ posed and worth increasing dollar values.
What we saw last week was impressive proof that many long-standing goals of Artificial Intelligence have been accomplished – namely that intelligent, non-human agents can do well at what have always been thought as only doable by humans. Specifically, Watson accomplished natural language processing and understanding, search of large bodies of textual information to find potential responses, reasoning over possible answers, probabilistic weighting of these options, phrasing of questions to the answers, and then human-sounding speech production of the best response.
So after 3 nights of competition against 2 all-time human champions, Watson won $77,147 – more than three times the next highest score of the two opponents, one with $24,000 and the other with $21,600. After watching the 3rd round, I thought proudly of my friends on the IBM DeepQA team, including the project lead, Dave Ferrucci, and fellow IBM researchers from my field – John Prager, Bran Boguraev, Eric Brown, Jennifer Chu-Carroll, Michael McCord, amongst others, whom I have worked both with and in competition against over many years of NLP and QA research. Hats off to them and their whole teams, for having accomplished that which many researchers such as myself and teams in labs like our Center for Natural Language Processing, have worked on and continue to work on for computer-supported intelligence and decision-making for applications in areas such as medicine, government, and law.
Again, congratulations to Watson, the DeepQA team, and all my fellow researchers in NLP and QA, no matter where you do your research. Our field has achieved a really big win!
====================================================================================
To hear more of Dean Liddy’s reaction to IBM’s Watson computer competing on Jeopardy, watch this video.
You can follow IBM Watson on twitter here.
Want to learn more about Watson? Here are several videos from the IBM YouTube channel explaining Watson’s capabilities:
IBM Watson: Computer Understands Natural Language
