Thursday, November 10, 2011

Watson

We have discussed some project failures so far in our course, which have been very telling, and even quite entertaining at times. I would like to share a brief chronicle of a highly successful IT project I have read about recently, IBM Watson.

Watson is a compute cluster built by the firm's DeepQA division. DeepQA's mandate is to perform research in the realm of artificial intelligence (AI) centring around human-like “open-domain” question answering. Open domain question answering is basically the phenomenon of a machine answering questions expressed in natural human language.

The 90 node, 2880 core, 80 teraFLOP Watson is among IBM's most impressive projects to date and in February, 2011, it beat the two top Jeopardy contestants of all time in a two game match demonstrating its capabilities.

This amazing feat was made possible, not only by brilliant engineers, architects, and powerful hardware, but highly organized and effective project managers as well. One of those project managers, Jim De Piante, sourced talent for practise competition against the supercomputer. Other key sub-projects included hardware delivery and staging, sample question development. Meanwhile, several other technical and non-technical project managers coordinated the efforts of at least nine other groups working in linguistics, systems, software development, game strategy (for playing Jeopardy), data linking, search, and more. Additionally, project managers had to coordinate all these complex sub-projects in order to complete the entire package that is Watson.

The budget was over three million dollars and the project was completed in approximately two years.

In order to ensure the project was completed on time, on specification, and on budget, project managers had to have a solid understanding of the ambitious project goals, resource constraints, software development, supercomputer hardware, human language, artificial intelligence, and working with a multicultural distributed team in four countries on a project some thought was impossible. It was an impressive feat indeed.



Sources:

http://www-03.ibm.com/innovation/us/watson/research-team/

http://en.wikipedia.org/wiki/Watson_(computer)

http://www-03.ibm.com/innovation/us/watson/

http://www-03.ibm.com/innovation/us/watson/building-watson/index.html

http://www.research.ibm.com/deepqa/deepqa.shtml