User Tools

Site Tools

Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/8/8b55c297879c7bc61a42ba1a61a09b95.i failed
Unable to save cache file. Hint: disk full; file permissions; safe_mode setting.
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/8/8b55c297879c7bc61a42ba1a61a09b95.metadata failed
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/4/4c47e099ad93cfdd49da21f4cec5aad6.i failed
Unable to save cache file. Hint: disk full; file permissions; safe_mode setting.
Writing /home/users/ashutosn/public_html/CommNetS2016/dokuwiki/data/cache/4/4c47e099ad93cfdd49da21f4cec5aad6.xhtml failed


This shows you the differences between two versions of the page.

Link to this comparison view

a_tensor_spectral_approach_to_learning_mixed_membership_community_models [2016/09/01 19:15] (current)
Line 1: Line 1:
 +Title: A Tensor Spectral Approach to Learning Mixed Membership Community Models
 +Abstract: Modeling community formation and detecting hidden
 +communities in networks is a well studied problem. However,
 +theoretical analysis of  community detection has been mostly limited
 +to models with non-overlapping communities such as the stochastic
 +block model. In this paper, we remove this restriction,​ and consider a
 +family of probabilistic network models with overlapping communities,​
 +termed as the mixed membership Dirichlet model, first introduced ​  in
 +Aioroldi et. al. 2008. This model allows for nodes to have fractional
 +memberships in multiple communities and assumes that the community
 +memberships are drawn from a Dirichlet distribution. We propose a
 +unified ​ approach to learning these models via a   ​tensor spectral
 +decomposition method. Our estimator is based on  low-order ​ moment
 +tensor of the observed network, consisting of  3-star counts. Our
 +learning method is fast and is based on   ​simple linear algebra
 +operations, e.g. singular value decomposition and tensor power
 +iterations. We provide guaranteed recovery of community memberships
 +and model parameters and present a careful finite sample analysis of
 +our learning method. Additionally,​ our results ​ match the best known
 +scaling requirements in the special case of the stochastic block
 +model. This is joint work with Rong Ge, Daniel Hsu and Sham Kakade.
 +Bio: Anima Anandkumar has been a faculty at the EECS Dept. at U.C.Irvine since Aug. 2010. Her current research interests are in the area of high-dimensional statistics and machine learning with a focus on learning probabilistic graphical models and latent variable models. She was recently a visiting faculty at Microsoft Research New England (April-Dec. 2012). ​ She was   a post-doctoral researcher at the Stochastic Systems Group at MIT (2009-2010). She received her B.Tech in Electrical Engineering from IIT Madras (2004) and her PhD from Cornell University (2009). She is the recipient of the ARO Young Investigator Award (2013), NSF CAREER Award (2013), ​ ACM Sigmetrics Best Paper Award (2011), ​ ACM Sigmetrics Best Thesis Award (2009), ​ IEEE Signal Processing Society Young Author Best Paper Award (2008), and  IBM Fran Allen PhD fellowship (2008). ​
a_tensor_spectral_approach_to_learning_mixed_membership_community_models.txt ยท Last modified: 2016/09/01 19:15 (external edit)