Tuesday, March 30, 2010

ACM Awards

ACM has announced some awards. Congratulations to:
  • Tim Roughgarden, recipient of the Grace Murray Hopper Award for introducing novel techniques that quantify lost efficiency with the uncoordinated behavior of network users who act in their own self-interest. His research has built a bridge between theoretical computer science and the networking research community that has the potential to capture the important role of strategic behavior in the design and analysis of future networks. Roughgarden is an assistant professor at Stanford University, whose book, Selfish Routing and the Price of Anarchy, outlines several approaches to limiting the efficiency loss in large networks resulting from self-interested users. The Hopper Award recognizes the outstanding young computer professional of the year.
  • Mihir Bellare and Phillip Rogaway, recipients of the Paris Kanellakis Theory and Practice Award for their development of practice-oriented provable security, which has resulted in high-quality, cost-effective cryptography, a key component for Internet security in an era of explosive growth in online transactions. Bellare, a professor at the University of California San Diego, and Rogaway, a professor at the University of California Davis, adapted modern cryptographic theory to make it more applicable for reducing the risk of cyber attacks in the real world. The Kanellakis Award honors specific theoretical accomplishments that significantly affect the practice of computing.



My research interest is fairly focused these days, mostly on problems in ad systems: algorithms, optimization, game theory, data mining, whatever. I pursue problems that I intuitively feel are relevant to the application, or new directions that I think should be developed by the community. It occurred to me, I could use a postdoctoral researcher to brainstorm ideas and solve problems with. May be it is late in the season to look for candidates. Anyone out there, interested in being at the Rutgers neighborhood and thinking about these exciting problems?


AdX: Industry view

Here is a whitepaper on Ad Exchanges, that gives an industry view:
The whitepaper does a nice job of laying out the landscape and explores some tricky topics of competition. Here is a set of blog-responses to the above: Darren Harmen, lookery, Ad Exchanger, Etc.

Btw, I will giving an Industry Track talk at the WWW 2010 conference on Ad Exchanges, Friday Apr 30, 1--2.30 PM. Any pointers, papers, etc. welcome. This talk will be about the design decisions in the product, engineering challenges, algorithmic and game-theoretic research issues, and finally, a discussion of the ecosystem (I dislike that word!) around Ad Exchanges.


Wednesday, March 24, 2010

Hack at NY

Chris Wiggins brought this 24 hour hackathon to my attention: First ever NYC-wide student hackathon, April 2-3, 7pm-7pm.
NYC-area Students:

Meet your fellow coders/developers/hackers from Columbia, NYU, CCNY, and other NYC area schools! Learn how to develop with APIs, datasets, and technologies demoed by NYC's hottest startups Improve your skills with tips+tricks from startup CTOs and your fellow students. Hackathon closes with student demos, competing to earn prizes and glory. Need more incentive? Free food and likely free Red Bull. Hopefully tshirts and startup swag.

Students: To attend RSVP to springhack@hackny.org. Include your full name, University, class year, and major. Visit http://hackNY.org or email info@hackny.org for more info.

about: HackNY.org is working to connect NYC students with NYC startups by facilitating a summer internship program and to build a community among the next generation of NYC hackers. Starting in 2010, the hackNY summer internship program includes pedagogical lectures as well as dinners with local and nonlocal upstarts, hackers, and founders.

Friday, March 19, 2010

Monitoring user feedback on the net

It takes a lot of talent to look at an applied situation, abstract a theoretical CS problem, solve it, and have it really matter in the application. You have to be creative and balance between the pull of the application to be ad hoc, and that of the theory to be principled and analyzable. If you go too far in any direction, you get speared. Mark Sandler pulled a nice example recently:

The problem is, there is a lot of user-generated content on the web. It may be inappropriate for many reasons. We don't have algorithmic and human resources to monitor all to see which is appropriate and which should be removed. So, one solution is to let users on the web monitor the content and provide feedback (keep or remove the content). Now the problem is the feedback may not be correct, and we need to monitor feedback! A solution is to carefully select a few of them to be vetted by human raters. What is a suitable formulation of this problem of trading off missing some of the inappropriate content vs using far too many human raters, for this task? Mark and I worked on this, tried many different formulations, but Mark came up with a very nice framework ultimately. We have this paper in the upcoming WWW conference. The paper leaves open some probabilistic analyses which could be fun.



On the book Snow, by Maxence Fermine. Yuko Akita chooses to be a poet (rather than a monk or a samurai) and writes haikus about snow. Video on the left has the rest.

Wednesday, March 17, 2010

The Storm before Spring

Today it is warm and trees are preparing for the Spring, but the past weekend saw a storm in the area ripping electricity lines, felling trees, and flooding basements.
  • I walked into the subway, wet and waited for the train. There was a Gypsi girl on the platform singing, you guessed it, Rihanna's Umbrella.
  • The city streets were not just puddles and racing cabs with glorious sprays, but a true collage of umbrellas, abandoned where they fell, twisted metal, collapsed shafts and ripped fabric.
  • What sorta Economy supports picking up an Umbrella when it rains and abandoning it before hopping into a cab? (cost of materials, labor for assembly, shipping, humans to distribute and sell it and feed their families, all for $5 total per Umbrella.)

Tuesday, March 16, 2010

Privacy Online

A NYT article with leads to bunch of analyses by researchers, more such technical leads than in most articles.

Friday, March 12, 2010

Ad Auctions Workshop 2010

The 6th Ad Auctions Workshop will take place on June 8th, collocated with the EC Conference. The call for participation is now available. Deadline April 14. CFP says "The workshop will bring together researchers and practitioners from academia and industry to discuss the latest developments in advertisement auctions and exchanges." I love how ad exchanges are now first class objects.

The organizing committee comprises: Moshe Babaioff, Ben Edelman, Jon Feldman, Sebastien Lahaie and Kamesh Munagala.

As Jon says, "Historically this is always a great workshop, more focused than the typical academic conference. It brings together folks from both industry and academia, and lots of different areas (algorithms, economics, optimization, machine learning)." And as CFP says, "The workshop’s proceedings can be considered non-archival, meaning contributors are free to publish their results later in archival journals or conferences." In other words, it is a true workshop.


Privacy strikes (again!)

Events unfold thus:
  • Netflix releases data for a stunning contest that captures the researchers' imagination.
  • Researchers correlate the data with publically available data to break anonymity. Netflix is sued.
  • Legalese, yaada, yaada, yaada ... Netflix cancels the version 2 of the contest.


Thursday, March 11, 2010

Ad Exchange on WSJ

Discussion about ad exchange on WSJ. Quote:
Advertisers say the new exchanges allow them to better target their advertising, though the auctions don't necessarily result in lower prices. In early testing, Google said its research found that the average price a publisher receives for ad space sold through its exchange is more than 130% higher than the average price of similar ad space sold directly to ad networks and other third parties.

Wednesday, March 10, 2010

Being a Writer

pauca sed magnus.

Indeed. Three examples of E. B. White:

Tuesday, March 09, 2010

EC 2010 Accepted Papers

Chris Dellarocas and Moshe Tennenholtz ran the PC and have released the list of accepted papers to EC 2010.

Saturday, March 06, 2010

Saturday Media on Media

When I travel in Italy, I watch TV shows that discuss the printed media of the day (so I imagine, I don't understand Italian, there is usually a host and a hostess and some aging pundit with bunch of newpapers, tabloids, spread out in front of him, and dramatic discussion). In that spirit, here is a video of me discussing a National Geographic article on the Hadza, a community of hunters and gatherers (yes, there are a few of those even in 21st century) in Tanzania.

Friday, March 05, 2010

Ad Exchange: Technical

I set up a site to collect technical + research material about ad exchanges. I will keep the site updated. For now, it has two new research papers:
  • Selective call-out and real time bidding. T. Chakraborty, E. Even-Dar, S. Guha, Y. Mansour and S. Muthukrishnan. ArXivs 2010. When publishers access the exchange, the exchange will call out ad networks selectively in real time if they have a bound on how many auctions they can participate. The paper studies two different pricing models, has some technical connection to buffer management in networks, uses primal-dual techniques for online approximation, and shows some experimental results. It is fair to say that the authors are excited about the way they have modeled this problem that is increasingly a challenge in practice.
  • Auctions with Intermediaries. J. Feldman, V. Mirrokni, S. Muthukrishnan and M. Pai. 2010. Myerson's theory says how to auction a good for optimal revenue. In the ad exchange, we have two levels of auctions with the exchange auctioning good to the networks and the networks in turn selling to their downstream advertisers. What is the optimal revenue auction for either party in this case? This paper presents optimal auctions for ad networks and for the exchange, there are some interesting twists to the theory.
More in the future.


Thursday, March 04, 2010

Sloan Awards

Alfred P. Sloan Fellowships have been announced for 2010. They had a one page ad in NY Times (A21). Congratulations to:
  • Constantinos Daskalakis
  • Jonathan Kelner
  • Amin Saberi
and others. Check out the full list across Economics, Math and others.