We present an efficient algorithm for learning mixed membership models when the number of variables p is much larger than the number of hidden components k. This algorithm reduces the computational complexity of state-of-the-art tensor methods, which require decomposing an O(p^3) tensor, to factorizing O (p/k) sub-tensors each of size O(k^3). In addition, we address the issue of negative entries in the empirical method of moments based estimators. We provide sufficient conditions under which our approach has provable guarantees. Our approach obtains competitive empirical results on both simulated and real data.
Zilong Tan (Duke University)
PhD student in Computer Science at Duke University. Interested in fast computational algorithms for scalable ML. Interned at Microsoft Research in Redmond and LinkedIn in Mountain View. Prior to Duke, I was a senior R&D engineer working on click-through-rate prediction at Baidu in Beijing.
Sayan Mukherjee (Duke University)
Related Events (a corresponding poster, oral, or spotlight)
2017 Talk: Partitioned Tensor Factorizations for Learning Mixed Membership Models »
Mon Aug 7th 03:48 -- 04:06 AM Room C4.4