Learning about Watson


At the Bring IT, Together Conference last week, I had slotted aside half a day to spend in Tim King’s session on security. As it would happen, unfortunately, the voicEd Radio show was being recorded at the time so I had to miss it.

I’m not sure whether or not presenters got a registrant list so I hunted him down over lunch to let him know of my absence. He didn’t seem to be too disappointed (maybe it was the birthday gift I’d given him last summer) but then he indicated that I needed to go to this other session in the afternoon. It was given by Gordon Alexander from IBM. He was going to talk about IBM’s Watson and augmented intelligence. He also had good IBM swag to give away. I like swag.

So, I went and thoroughly enjoyed things. While Mr. Alexander had to step aside for a couple of minutes at the start, Tim begin the session and, of course, we had to get up to speed with Watson and its work on Jeopardy.

I’d seen it before but was interested all the same. Then, Mr. Alexander came back and delivered the message that was the heart of the presentation. Then, we got a chance to try it ourselves.

All of us in the group were invited to create a free account on Watson and it was smart enough to reject this flood of requests from the same location! So, we switched off the wifi on our phones and created the accounts on our own data. Success.

The activity that were to take part of was to create our own chatbot. Tim’s TEJ class had had the same presentation as we did and they did things like an interactive pizza ordering bot. That would definitely be of interest to students and so we were off. The interface was very much like flowcharting from years ago before flowcharting kind of went away for programming as we shifted away from a procedural paradigm. It seemed very natural and fluent for me and I was plugging away when Tim offered an already created product that he thought we should explore. Diversion time.

It was called Personality Insights and the claim was that Watson would take your content from Twitter and make observations about you as a user. I guess it makes sense since when you post to Twitter, you don’t have any expectation of privacy. So, we gave it a shot. We also found out that the source is on GitHub.

By this time, Tim’s wife Alanna had come along side of me. She wanted to see my results. Only if I could see hers! She grabbed her tablet and off the two of us went off to explore.

Thankfully, I had a fully qualified teacher-librarian sitting beside me to explain some of the things that got analyzed. Gregariousness? It was interesting. A couple of people had tweeted out the direct link and others who saw it were checking it out. It’s too bad because they would have lost the context of being in the workshop.

But, the bigger thing was to go to Watson’s page and the personality insights page to get the bigger picture of the program/product. It added a great deal to what I think about when I talk about advertising and business. Overtly, who hasn’t chatted with a bot on a car dealership’s website? What is happening that we don’t know about?

Very quickly, the 2.5 hours was over. We hadn’t had a break because we were so engaged with Watson and the discussions surrounding it. I found the conversation about guidance departments and student support with products like Watson very interesting. Maybe even a bit scary. What happens if Watson is wrong?

So, thanks, Tim for steering me in this direction for that afternoon. I really enjoyed it.

And, I got a pair of blue IBM ear buds to take home with me. I wonder if they’re “powered by Watson”?

OTR Links 11/12/2019


Posted from Diigo. The rest of my favorite links are here.