Well, June is conferences season for me, so despite a new baby at home I went off on another trip this week — sorry honey! This time it was ACM Sigmetrics in Austin, where I helped to organize the GreenMetrics workshop, and then presented one of our group’s three papers on the first day of the main conference.
Here’s the full list of the the three from our group – you can see that they cover a range of applications, but they all have some form of pricing/markets involved:
- Energy Procurement Strategies in the Presence of Intermittent Sources Jayakrishnan Nair (CWI); Sachin Adlakha (California Institute of Technology); Adam Wierman (California Institute of Technology)
- Pricing Data Center Demand Response Zhenhua Liu (California Institute of Technology); Iris Liu (California Institute of Technology); Steven H. Low (California Institute of Technology); Adam Wierman (California Institute of Technology)
- Pricing Link by Time Chengdi Lai (The University of Hong Kong); Steven H. Low (California Institute of Technology); Ka-Cheong Leung (The University of Hong Kong); Victor O.K. Li (The University of Hong Kong)
I presented the energy procurement paper, and I was surprised at how foreign it felt to prepare a conference talk! I realized students/postdocs have been responsible for presenting all my conference papers recently, so it had been five years (I think) since I’d given a conference presentation anywhere other than INFORMS, which is a very different beast. Being so used to the hour-seminar style, it took me a bit of work to actually get things down to the 20min limit, but I think the result came out well. You can see the slides here, with the normal disclaimers that I tend to have sparse slides and say a lot more than is actually on them…
Growing interest and changing topics
I think the big headlines from the conference this year were about growth. Sigmetrics had record submissions, a record number of acceptances, and will likely have record attendance after the on-site registrations are included. I think big congratulations should go to the organizers for this – Sujay Sanghavi, Sanjay Shakkottai, Bianca Schroeder, and Marc Lelarge. I think that’s a great sign for the conference, and it was especially nice to see the conference growing in directions that haven’t historically been well-represented like economics, social networks, big data, and power.
Another thing that stood out was that I saw a lot of the same folks that were at EC at Sigmetrics. I’d say the overlap was at least 15 folks, and those are the ones I knew, so there was likely more. It’s nice to know that our group isn’t alone in straddling those boundaries, and I think a lot of the overlap has come from the colocation of the NetEcon workshop with Sigmetrics these last few years. And, given EC and Sigmetrics will be collocated at the FCRC next year, I think there’ll be great opportunities to build connections between the communities even further.
Some highlights from days 1 & 2
Maybe the biggest highlight of the conference for me were the two major award winners. Francois Baccelli won the Achievement Award, and Florian Simatos received the Rising Star Award. In both cases, these are researchers that have made foundational contributions to applied probability, so it’s great to see recognition of their work in the Sigmetrics community. Further, in Francois’ case, even though I haven’t collaborated with him, his work has been an inspiration for me throughout my career. He has pushed to develop theories (on multiple occasions) that were not accepted as “practical” right away, but their practical importance came to be recognized in big ways years down the road. He’s proven time and again that he has just amazing taste for understanding what is important in a model, and a vision for where mathematical tools can provide more than just a new algorithm/protocol; they can provide an entire new approach for reasoning about design.
Among the papers themselves, I thought the quality was extremely high (and the level of the presentations was also quite good). It makes it hard for me to pull out just a few to highlight here, so instead, I’ll highlight a couple themes that I was quite pleasantly surprised to see in the papers. There were multiple papers that were able to adapt recent results in approximate mechanism design (which is one of the biggest contributions of the Algorithmic Game Theory community in recent years, in my opinion) to practical problems, specifically cloud computing markets and electricity markets.
- An Online Auction Framework for Dynamic Resource Provisioning in Cloud Computing Weijie Shi (The University of Hong Kong); Linquan Zhang (University of Calgary); Chuan Wu (The University of Hong Kong); Zongpeng Li (University of Calgary); Francis C. M. Lau (The University of Hong Kong)
- Randomized Auction Design for Electricity Markets between Grids and Microgrids Linquan Zhang (University of Calgary); Zongpeng Li (University of Calgar); Chuan Wu (The University of Hong Kong)
These two definitely fell into the category of “I wish I had had that idea” papers… It’s really nice to see ideas transfer between EC and Sigmetrics communities in this way.
The transfer also seems to be happening in the domain of privacy. In my post about EC, I highlighted a paper on privacy-preserving recommendation systems, and again at Sigmetrics, there was a very nice paper that looked at the same issue, in a different (slightly more applied) perspective.
- Privacy Tradeoffs in Predictive Analytics Stratis Ioannidis (Technicolor); Andrea Montanari (Stanford); Udi Weinsberg (Technicolor); Smriti Bhagat (Technicolor); Nadia Fawaz (Technicolor); Nina Taft (Technicolor)
Definitely worth a read by folks interested in privacy that weren’t at Sigmetrics.