Ganesh Swami

Not many people seem to know about the WWW07 conference, so I’m sending them some love. Defined by the organizers as “the global event that brings together the key innovators, decision-makers, technologists, businesses, and standards bodies shaping the Web,” the conference has some interesting refereed papers and posters.

The Web space as of now is pretty crowed and it’s difficult to differentiate the wheat from the chaff. The next social network, or the next youtube clone isn’t very interesting. Skimming through these papers brings to light some of the harder problems in the field:

• spammers posting portions of existing comments with a few links changed.
• a website that has the authority in one area might start writing about other things. How do you assign a pagerank to this site?
• RSS aggregators clustering. No fun reading about five different takes on the same breaking news. Much like Google News.
• Google news itself. “collaborative filtering using MinHash clustering, Probabilistic Latent Semantic Indexing (PLSI), and covisitation counts.”
• mining and clustering other kinds of information (like chemical formulae)
• advertisements and click fraud
• personalization
• and a whole bunch of stuff on scalability, privacy, security and the semantic web

Also speaking at the event is Prabhakar Raghavan. I have previously heard him talk about the convergence of the social sciences and the web. Awesome speaker.

I came across a very cool project to rank academic journals: EigenFactor.org. Quoting the about page:

Eigenfactor.org is a non-commercial academic research project sponsored by the Bergstrom lab in the Department of Biology at the University of Washington. We aim to develop novel methods for evaluating the influence of scholarly periodicals and for mapping the structure of academic research. We are committed to sharing our findings with interested members of the public, including librarians, journal editors, publishers, and authors of scholarly articles.

The ranking is done using an algorithm similar to Google’s PageRank. Read about the other features on their “Why EigenRank?” page.

Here for example is a listing of the highest ranking (by eigenfactor) journals in Mathematical Physics:

1. PHYSICAL REVIEW E
2. JOURNAL OF PHYSICS A-MATHEMATICAL AND GE
3. COMMUNICATIONS IN MATHEMATICAL PHYSICS
4. JOURNAL OF COMPUTATIONAL PHYSICS
5. JOURNAL OF MATHEMATICAL PHYSICS
6. PHYSICA D-NONLINEAR PHENOMENA
7. JOURNAL OF STATISTICAL PHYSICS
8. COMPUTER PHYSICS COMMUNICATIONS
9. MODERN PHYSICS LETTERS A
10. CHAOS

I’m yet to figure out a way to make use of this service effectively. My day job deals with Medical Imaging, which is a form of Computer Vision, which in turn falls under Artificial Intelligence. Computational Anatomy, one of the emerging subfields of medical imaging is fairly mathematical, using ideas from group theory and functional analysis. To add more to the mix, one of the most used algorithms for image/volume segmentation is the Level Set Method, first published in the Journal of Computational Physics. I can’t judge a paper just by looking at where it’s published, can I?

Through bbgm, I was introduced to the idea to Tumblelogs. I really like the format that is used to display content. I think limiting one post per page really helps focus the reading, and it doesn’t necessarily have to be short posts. I’m going to adopt this format for my blog over the next few months. Unfortunately, I’m currently swamped with work, so it’ll take atleast a month to see any changes on this site.