Majid Janzamin

Spectral Learning on Matrices and Tensors

Apr 16, 2020
Majid Janzamin, Rong Ge, Jean Kossaifi, Anima Anandkumar

Figure 2 for Spectral Learning on Matrices and Tensors

Figure 3 for Spectral Learning on Matrices and Tensors

Figure 4 for Spectral Learning on Matrices and Tensors

* Foundations and Trends in Machine Learning: Vol. 12: No. 5-6, pp 393-536 (2019)

Access Paper or Ask Questions

Provable Tensor Methods for Learning Mixtures of Generalized Linear Models

Jan 13, 2016
Hanie Sedghi, Majid Janzamin, Anima Anandkumar

* To appear in Proceeding of AI and Statistics (AISTATS) 2016

Access Paper or Ask Questions

Beating the Perils of Non-Convexity: Guaranteed Training of Neural Networks using Tensor Methods

Jan 12, 2016
Majid Janzamin, Hanie Sedghi, Anima Anandkumar

Figure 2 for Beating the Perils of Non-Convexity: Guaranteed Training of Neural Networks using Tensor Methods

* The tensor decomposition analysis is expanded, and the analysis of ridge regression is added for recovering the parameters of last layer of neural network

Access Paper or Ask Questions

Analyzing Tensor Power Method Dynamics in Overcomplete Regime

Sep 14, 2015
Anima Anandkumar, Rong Ge, Majid Janzamin

Figure 2 for Analyzing Tensor Power Method Dynamics in Overcomplete Regime

Figure 3 for Analyzing Tensor Power Method Dynamics in Overcomplete Regime

* 38 pages; analysis of noise added to the previous version

Access Paper or Ask Questions

Score Function Features for Discriminative Learning

Apr 19, 2015
Majid Janzamin, Hanie Sedghi, Anima Anandkumar

* Accepted as a workshop contribution at ICLR 2015. A longer version of this work is also available on arXiv: http://arxiv.org/abs/1412.2863

Access Paper or Ask Questions

Guaranteed Non-Orthogonal Tensor Decomposition via Alternating Rank-$1$ Updates

Mar 04, 2015
Animashree Anandkumar, Rong Ge, Majid Janzamin

Figure 2 for Guaranteed Non-Orthogonal Tensor Decomposition via Alternating Rank-$1$ Updates

Figure 3 for Guaranteed Non-Orthogonal Tensor Decomposition via Alternating Rank-$1$ Updates

* We have added an additional sub-algorithm to remove the (approximate) residual error left after the tensor power iteration

Access Paper or Ask Questions

Sample Complexity Analysis for Learning Overcomplete Latent Variable Models through Tensor Methods

Dec 16, 2014
Animashree Anandkumar, Rong Ge, Majid Janzamin

Figure 2 for Sample Complexity Analysis for Learning Overcomplete Latent Variable Models through Tensor Methods

Figure 3 for Sample Complexity Analysis for Learning Overcomplete Latent Variable Models through Tensor Methods

Figure 4 for Sample Complexity Analysis for Learning Overcomplete Latent Variable Models through Tensor Methods

* Title changed

Access Paper or Ask Questions

Score Function Features for Discriminative Learning: Matrix and Tensor Framework

Dec 11, 2014
Majid Janzamin, Hanie Sedghi, Anima Anandkumar

* 29 pages

Access Paper or Ask Questions

High-Dimensional Covariance Decomposition into Sparse Markov and Independence Models

Dec 14, 2013
Majid Janzamin, Animashree Anandkumar

Figure 2 for High-Dimensional Covariance Decomposition into Sparse Markov and Independence Models

Figure 3 for High-Dimensional Covariance Decomposition into Sparse Markov and Independence Models

Figure 4 for High-Dimensional Covariance Decomposition into Sparse Markov and Independence Models

* 42 pages. Experiments are updated, and the optimization program is implemented by an ADMM technique which makes it faster

Access Paper or Ask Questions

When are Overcomplete Topic Models Identifiable? Uniqueness of Tensor Tucker Decompositions with Structured Sparsity

Aug 13, 2013
Animashree Anandkumar, Daniel Hsu, Majid Janzamin, Sham Kakade

Access Paper or Ask Questions

1

2

>>