Building Watson – A Brief Overview of the DeepQA Project

David Ferrucci of IBM discusses the DeepQA Project; the technology and architecture behind IBM’s newest technological innovation, the question answering and natural language processing system, Watson.

Visit for more information.

40 thoughts on “Building Watson – A Brief Overview of the DeepQA Project

  1. @jameschuston
    another typo in your correction itself.

    ln ((12546798*pi) )^ 2 / 34567.46 = 0.00885 (instead of your 0.00855 🙂

    and removing the extra parens.. simplifies to

    ln (12546798*pi) ^ 2 / 34567.46 = 0.00885


  2. @AnitaCock of course. please make any other scientific breakthrough completely insignificant, which was made possible with the help of IBM because I beg you to overrate the finding you think you have made there even more. How tasty was my little frenchman?

  3. Actually, the answer to the biology question is wrong too.

    The LIQUID that "cushions" the nucleus is the cytosol, not the cytoplasm. The cytoplasm is the cytosol plus non-liquid parts such as mitochondria, lysosomes, ribosomes, cytoskeletal elements, proteins, and so on.

    If the question asked for the LIQUID, the answer would be cytosol, not cytoplasm. A fluid is either a gas or a liquid, so even expanded it to fluid instead of liquid doesn't change the fact that cytosol is the better answer

  4. I imaging my 3 year old son telling me sooner or later: "what is the point of learning all that stuff in the school, if computer will find the answer a whole lot quicker?".

  5. Part 1:
    I'm hoping they're working on a front-end that you can have a conversation with in the process..the FUN is going to be when you're asking a question..the computer asks a question back..then you go back and forth to figure out how to get the answer ..the BAD thing is i HOPE they don't go so far down the path that it'll block that from happening because in MY opinion in order to have a conversational computer it'd have to have a SMALLER database on the front for faster access ..

  6. Part 2:
    ..because when I think about the way "i" think ..if somebody says "hey how's it going?"..I don't think about aLLLL the possibilities..i say " so" (..been a boring year) ..AND the reason i say 'so so' is because i've heard someone else say that recently etc. (it seems like memory is also laid out pretty deeply according time)..AND..if i tried to think of EVERY perfect's take me..about 2 hours toO ..i guess that's the term DeepQA though!..
    Can't wait for ShallowQA!!

  7. If I could ask Watson something it would be: What is the meaning of life? and I wonder if it would answer something like… "the meaning of life if to help the others". Can Watson give me not just a right answer but a wise answer?

  8. If IBM is gonna make a even better computer than Watson, they should call it Sherlock Holmes. So it would be Sherlock Holmes and Watson :)

  9. But Incredible i hear many many people talking about wrong concept "programmed = predictable" , but you can program , a self programmable system ( with learn capacity ) , after watching Watson sounds obvious ? …

  10. bloody hell you have such a great technology… and what u doing with it….
    make it open source. let ppl find its applications.

  11. Watson can be interfaced with other communications type technologies to include voice and text based communications.

    But even at current level if it does not understand the question would it not ask for clarity or if there is an index of common questions like this it could provide references to such questions.


  12. Watson follows academics because you have specialist who can only understand medical literature, research and publication. Doctors like me who work in intensive and critical care for long time have brain probably working faster than Watson. This is called "Clinical acumen" and I have incorporated this in a simple tool called "MAYA (Medical Advice You Access)" will any day give you better advice than this computer can.

  13. Academics working in the NHS in UK started working on the hypothesis of standardising healthcare in 1995. Initially they introduced "Pre-printed Questioners", then brought in "Evidence based medicine". In 2003, allowed nurses to offer treatment based on protocols and algorithms. The result of this callous attitude 37,000 people died due to septicaemia. Offering healthcare advice & treatment is not based only on knowledge but being human 

  14. Well If its so super.. its should be able to decode this…

    01010100 01101000 01100101 00100000 01101101 01100101 01100001 01101110 01101001 01101110 01100111 00100000 01101111 01100110 00100000 01101100 01101001 01100110 01100101 00100000 01101001 01110011 00100000 01110100 01101111 00100000 01101000 01100101 01101100 01110000 00100000 01101111 01110100 01101000 01100101 01110010 01110011

  15. Sounds and looks great but… will there be ethics in its applications? It will always be a machine. No feelings in the sense of metanoia capability. Will there ever be an algorithm for love and hate? More importantly an algorithm to help it choose love over hate? Because that's where all this is headed sooner or later mark these words.

  16. How does this thing get new information? What if you asked a Jeopardy question about something that happened today? It's not connected to the internet, so how is the database updated?

  17. Hi, How many kind algorithms used to develop the Watson.  Its good to understand the application of Watson in all subjects…. But still  the world of technology unaware with core technology of #Watson.

Leave a Reply