I  attended this year's ACM KDD conference from 8/12-8/16. This is the first time I went to the conference even though I've been working on data mining for years. I attended IEEE ICDM and SIAM ICDM before.

I like it this year in general. However, like in other similar conferences, I only found several interesting presentations and the rest kind of boring. On sunday I found Pedro Domingos's tutorial on Markov Logic Netowrks to be interesting and informative. I like the invited talks given by Usama Fayyad and Chris Anderson.  The video recordings of many of these talks are actually available on videolectures.net. In terms of technical papers, I found the best paper award winners (and runners-up)  to be very good. The industry track was in parallel and arranged in a room that is on a differen floor as the other sessions. I didn't find time to go any of the industrial track sessions. The best thing I like is, of course, to meet with lots of friends, colleagues, and renowned data mining researchers from all over the world.

As usual, a lot of companies were there making a presence and hiring people. Microsoft, Google, Yahoo, SAS, Oracle, SPSS, Fair Isaac, etc.