Research Scientist, Eric Bax
I had a fun visit to the UC Irvine computer science department on February 11, 2013. It was great to see professor Padhraic Smyth again after meeting in a NIPS workshop and to meet professor Alex Ihler. John Turner, a professor from the business school, joined in as well, contributing his acumen to the hands-on activity in my talk.
Some student teams were advertising companies and others were data provider companies. The challenge for advertisers was to place winning bids for the best ad calls. The challenge for data providers was to profit from selling their data about which ad calls were most valuable.
It was great fun — a very active group. There was some pretty intense negotiation and deal-making: exclusive data rights, deals for percentages of profits, the back-and-forth of agreeing on prices. Once the ad auction was held, we had a chance to analyze which teams made the most money and why.
We had a lot of fun with an interesting problem — setting prices for information. It is a neat problem because it touches many areas of study, including computer science, statistics, economics, business, and public policy.
The students had a great discussion about the activity, and they came up with interesting ideas on a number of fronts: negotiation, deal structure, and thinking about the nature of the problem itself.
After the talk and activity, I met with some students. One, Chris DuBois, is doing interesting work on networks that evolve over time. I connected him with James Li, at Cornell, who is doing work on similar problems. James was an intern here at Yahoo a few summers ago, and we still collaborate on follow-on research.