A
Ankur Moitra
Researcher at Massachusetts Institute of Technology
Publications - 166
Citations - 6970
Ankur Moitra is an academic researcher from Massachusetts Institute of Technology. The author has contributed to research in topics: Estimator & Polynomial. The author has an hindex of 42, co-authored 153 publications receiving 6024 citations. Previous affiliations of Ankur Moitra include Institute for Advanced Study.
Papers
More filters
Proceedings Article
A Practical Algorithm for Topic Modeling with Provable Guarantees
Sanjeev Arora,Rong Ge,Yoni Halpern,David Mimno,Ankur Moitra,David Sontag,Yichen Wu,Michael Zhu +7 more
TL;DR: This article presented an algorithm for learning topic models that is both provable and practical, which produces results comparable to the best MCMC implementations while running orders of magnitude faster than MCMC.
Proceedings ArticleDOI
Computing a nonnegative matrix factorization -- provably
TL;DR: This work gives an algorithm that runs in time polynomial in n, m and r under the separablity condition identified by Donoho and Stodden in 2003, and is the firstPolynomial-time algorithm that provably works under a non-trivial condition on the input matrix.
Posted Content
Learning Topic Models - Going beyond SVD
TL;DR: In this article, the authors formally justify nonnegative matrix factorization (NMF) as a main tool in this context, which is an analog of SVD where all vectors are nonnegative.
Proceedings ArticleDOI
Learning Topic Models -- Going beyond SVD
TL;DR: This paper formally justifies Nonnegative Matrix Factorization (NMF) as a main tool in this context, which is an analog of SVD where all vectors are nonnegative, and gives the first polynomial-time algorithm for learning topic models without the above two limitations.
Posted Content
Settling the Polynomial Learnability of Mixtures of Gaussians
Ankur Moitra,Gregory Valiant +1 more
TL;DR: In this article, an efficient algorithm for learning univariate mixtures of two Gaussians was proposed, which has a running time polynomial in the dimension and the inverse of the desired accuracy, with provably minimal assumptions on the Gaussian parameters.