Recently I attended a meetup at the IBM Silicon Valley Labs in South San Jose. The title of the meetup was IBM Watson to enable new generation of apps to solve business/social problems. Because it was a good source of information, I’ve summarized it briefly. Also, take a look at this, which gives more information on Watson from Robert High.
There were four speakers:
- Robert High, IBM Fellow, Vice President and CTO
- Dr. Jeffrey Kreulen, Distinguished Engineer, Watson Architecture, IBM Watson
- Rama Akkiraju, Distinguished Engineer, Personality Insights, IBM Watson
- Dr. Shivakumar Vaithyanathan, Fellow, Watson Cognitive Services, IBM Watson
Overview and Summary
High gave an overview, and a summary at the end. He also entertained questions after the talks.
He started with Watson’s history. Everyone knows how Watson beat two contestants at Jeopardy in 2011. A lot of media coverage and videos are available on YouTube and elsewhere. High wanted to make the point that Watson needed media attention, and he admitted that its participation in the show was a marketing stunt. Now Watson’s name is firmly tied to Jeopardy, but what’s going on with Watson four years after the Jeopardy victory? High listed the highlights of its progression year by year, as shown below. Because I could not find an overall progress list like this easily on the net, I found it very useful.
Rob High and Watson’s progression year by year
It is noteworthy that the first vertical market segment it jumped into was healthcare. High mentioned that advancements in the medical field are astonishing. To keep up with vast amounts of ever-changing information, each doctor must absorb a great deal of new information constantly, whether it is in the form of media articles or of papers in medical journals or at conferences. Doctors, especially specialists, are notoriously busy. How can they find time to locate the appropriate sources of information for their expertise, let alone digest such a massive amount of information?
High touched on other Watson offerings (shown on the IBM Watson web page), including the ones on High’s list above (Discovery Advisor and Explorer). In 2014, Watson Ecosystem was established, making it possible to collaborate with other organizations to use their results and vice versa. In 2015, a set of APIs was introduced on Bluemix (PaaS cloud platform). Bluemix is a commercial version of Cloud Foundry (open source).
APIs for Watson functions are listed below:
Watson APIs available on Bluemix (shown on the Bluemix page)
Note the following links, if you plan to play with Watson APIs on Bluemix. I created the list and am sure I missed some important ones. Let me know if you want to add any.
- Bluemix Developer Community
- Top 38 Questions about the Watson Services on Bluemix
- Watson blog
- Some whitepapers
- IBM Watson Meetup (This is for the San Francisco Bay Area. Check your area for similar meetups.)
- Twitter: @IBMWATSON
A demo using Watson APIs on Bluemix is available.
- Personality Insights, which was discussed by the third speaker.
There are a couple of other demos that will be discussed in Part 2 of this blog.
I just want to add that IBM is expanding Watson services outside the United States. In February 2015, IBM and Softbank in Japan (owner of Sprint) announced an alliance for the Japanese market. (In November 2015, IBM announced its expansion to the Netherlands. Although I do not speak Dutch, I am sure Dutch is closer to English as a language.) In Japan, Softbank takes on the role of sales and marketing for Watson, while IBM and IBM Japan work on the Japanizing of Watson. According to Softbank, the Japanese version of Watson is available in 2015, but most Watson APIs have not been Japanized. I wonder how much needs to be changed to run Watson in Japanese? Softbank wanted to combine Watson’s brain and its humanoid robot Pepper. The initial 1,000 units were sold in a matter of minutes.
High touched a little bit on what cognitive computing is, which the next speaker covered in more detail.
Continued to Part 2.