Saturday, August 03, 2013

ACM COSN Update

ACM COSN program is out. Oct 7--8 in Boston.

The keynote is by the incomparable Jaron Lanier. The PC had to meet a very aggressive schedule and they did, thanks to the PC and PC Chairs (Amr and Bala) . 22 papers have been accepted for a very low acceptance rate.  Bala is quarterbacking the conference, all the way.

The PC process was less "social" than I imagined (wonder if we should have used Twitter or Google+, :)).  I would like to see a nudge in the future towards more theoretical papers (more submissions please), with less emphasis on hardcore empirical analyses. We have to keep in mind that these online social systems and phenomena we study are built by engineers and change frequently, so what endures is theoretical insight, novelty and boldness. Nevertheless, the program is interesting and broad, I am sure the community will find it exciting and useful to be at the conference.  This is the first ACM conf on Online Social Networks, so attend and be part of the history! 

Labels:

3 Comments:

Blogger Luca Aceto said...

Muthu,

Would you recommend "Who Owns the Future"? The reviews make it sound like an excellent read.

Cheers,

Luca

4:09 AM  
Anonymous Alex Lopez-Ortiz said...

We have to keep in mind that these online social systems and phenomena we study are built by engineers and change frequently, so what endures is theoretical insight, novelty and boldness.

Sure, but we still need data to base our theoretical models upon. E.g. is the graph sparse? what is the median degree or diameter? and most importantly "do any of these properties seem to be invariant?"

Otherwise what good is, say a proof of NP completeness on a general graph if as it turns out is polynomial on graphs of average low degree, or alternately a fast algorithm which assumes all nodes are low degree when many social networks have at least a few highly connected nodes?

I think the conference needs both sides of the coin. Empirical data that provides insight into the problem, and algorithms and models that hold regardless of day-to-day variations on the actual data.

2:31 PM  
Anonymous Anonymous said...

Hi Luca,
Yes.

Hi Alex,
We should some time. I dont buy the modelling argument (study data, design model, design methods, change/understand the world). It applies in some areas (Physics, Universe, Bio) and not in others to the same extent, and I have a large rant about its application to Internet systems.

-- Metoo

4:59 PM  

Post a Comment

<< Home