Kannan Achan Post doctoral researcher Microsoft Research - Silicon valley	Kannan Achan is currently a Post doctoral researcher at Microsoft Research Silicon valley. Prior to this, he was a member of the Probabilistic and Statistical Inference Group (PSI) at the University of Toronto. He received his PhD degree in Computer Science from the University of Toronto, Canada, under the supervision of Brendan Frey and Sam Roweis. He also has a graduate degree in mathematics. Kannan’s research interests are in machine learning, graphical models and their applications to statistical signal processing. More generally, his interests are in the design and implementation of algorithms and applications that learn from data. Curriculum Vitae
Mailing Address: Microsoft Corporation 1065 La Avenida Mountain view CA 94043 Phone:650 693 3895 Email:kachan@microsoft.com kachan@ieee.org
Current projects at MSR include: A (distributed) machine learning toolkit using DryadLINQ – with Mihai Budiu Modeling network level properties of end user hosts with the goal of (almost) content independent spam detection/prevention and botnet tagging -- with Yinglian Xie and Fang Yu Inference in graphical models with arbitrarily continuous random variables – with Michael Isard and John MacCormick. Applications of Prototype driven learning to natural language processing and other domains -- with Moises Goldszmidt and Lev Ratinov Inference in Click graph – with Ariel Fuxman and Panayiotis Tsaparas Selected publications: Using the Wisdom of the Crowds for Keyword Generation Ariel Fuxman, Panayiotis Tsaparas, Kannan Achan, Rakesh Agrawal. In Proceedings of WWW 2008 (to appear). How Dynamic are IP Addresses? [PDF] Yinglian Xie, Fang Yu, Kannan Achan, Eliot Gillum, Moises Goldszmidt, Ted Wobber. In Proceedings of ACM SIGCOMM 2007, Kyoto, Japan. A Segment-based Probabilistic Generative Model of Speech [PS.gz] [PDF] Kannan Achan, Sam T. Roweis, Aaron Hertzmann, Brendan J. Frey, IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP) 2005 Variational Speech Separation of More Sources than Mixtures [PS.gz | PDF] Steven Rennie, Kannan Achan, Parham Aarabi, Brendan J. Frey, Artificial Intelligence and Statistics (AISTATS 2005).   Learning to Cluster using Local Neighborhood Structure [PS.gz | PDF] Rómer Rosales, Kannan Achan, Brendan J. Frey, International Conference on Machine Learning 21 (ICML), Jul 2004 Probabilistic Inference of Speech Signals from Phaseless Spectrograms [PS.gz | PDF] Kannan Achan, Sam T. Roweis, Brendan J. Frey, Advances in Neural Information Processing Systems 16 (NIPS). Unsupervised Image Translation [PS.gz | PDF] Rómer Rosales, Kannan Achan, and Brendan Frey, IEEE International Conference on Computer Vision (ICCV), Nice, France. Oct 2003. A factorized variational technique for phase unwrapping in Markov random fields [PS.gz | PDF] Kannan Achan, Brendan. J. Frey, Ralf Koetter Uncertainty in Artificial Intelligence (UAI), 2001. Use of Tabu Search in the logical topology design of wavelength routed optical networks Kannan Achan, Yash Aneja and Subir Bandyopadhyay, High Performance Computing Systems and Applications, Kluwer Academic Press, 2001 Recent patents: Clustering Botnet behavior using parameterized models, pending, filed 2008 Kannan Achan, Yinglian Xie, and Fang Yu Query classification based on query click logs, pending, filed 2008 Ariel Fuxman, Panayiotis Tsaparas, Kannan Achan and Rakesh Agrawal Automatic Botnet spam signature generation, pending, filed 2008 Yinglian Xie, Fang Yu, Kannan Achan, Rina Panigrahy, Ivan Osipkov and Geoff Hulten Adding prototype information into probabilistic models, pending, filed 2007 Kannan Achan, Moises Goldszmidt and Lev Ratinov Automatically identifying dynamic internet protocol addresses, pending, filed 2007 Kannan Achan, Eliot Gillum, Yinglian Xie and Fang Yu
Previous affiliations: psi@toronto learning@toronto cs@uwaterloo

Thanks to Andrew Birrell for making the (Celtic) decorative bars available online.

Tuesday, December 04, 2007 1:26:06 AM

Wednesday, February 27, 2008 11:28:26 PM

Thursday, March 20, 2008 7:42:15 PM

Monday, April 21, 2008 5:11:14 PM