Tag Archives: low-rank

Divide-and-Conquer Learning by Anchoring a Conical Hull

Many well-known machine learning methods aim to draw a line between two classes. However, in our recently accepted NIPS 2014 paper “Divide-and-Conquer Learning by Anchoring a Conical Hull“, we reduce lots of fundamental machine learning problems (a broad class of … Continue reading

Posted in Tianyi's work | Tagged , , , , , , | Leave a comment

AISTATS 2013 GreBsmo code is released

Here is the GreBsmo code for our AISTATS 2013 paper. You can use it as a greedy version of GoDec solver for X=L+S problem. It is much faster and more robust. There are three video subsequences you can play in … Continue reading

Posted in Tianyi's work | Tagged , , , , | Leave a comment

[Best student paper award] Welcome to my “Divide-and-Conquer Anchoring (DCA)” talk at ICDM Dallas Dec 8

Is it possible to finish a 60000×10000 matrix decomposition (NMF, PCA, etc)  or completion in 6 seconds on your laptop’s matlab? Can we make it even faster by a simple distributable scheme? How to summarize a huge-scale dataset (ratings, movie, … Continue reading

Posted in Tianyi's work | Tagged , , , , , , | Leave a comment

Greedy Bilateral (GreB) Paradigm for Large-scale Matrix Completion, Robust PCA and Low-rank Approximation

Our paper “Greedy Bilateral Sketch, Completion and Smoothing” has been accepted by AISIATS 2013. Abstracts reads below, PDF is here, and code will be coming soon. Abstract: Recovering a large low-rank matrix from highly corrupted, incomplete or sparse outlier overwhelmed … Continue reading

Posted in Tianyi's work | Tagged , , , , , | 3 Comments

Semi-Soft GoDec: >4 times faster, auto-determined k

Here is a good news of GoDec (pertaining to our ICML 2011 paper): Semi-Soft GoDec is released. Different from the ordinary GoDec which imposes hard threshholding to both the singular values of the low-rank part L and the entries of the … Continue reading

Posted in Tianyi's work | Tagged , , , , , , , | 2 Comments

News about GoDec code and ICML 2011 paper

We recently published a google site for Go Decomposition (GoDec), presented on ICML 2011. Now you can find all the available information and upcoming news about GoDec on http://sites.google.com/site/godecomposition  On the new site, there are 3 resources about GoDec we … Continue reading

Posted in Tianyi's work | Tagged , , , , , , , | 5 Comments

GoDec: Randomized Low-rank & Sparse Matrix Decomposition in Noisy Case

This paper is accepted by ICML 2011 for presentation. Now the final version is ready and can be downloaded from here GO. Abstract: Low-rank and sparse structures have been profoundly studied in matrix completion and compressed sensing. In this paper, we develop … Continue reading

Posted in Tianyi's work | Tagged , , , , , , , , | 2 Comments

Multi-label Learning via Structured Decomposition and Group Sparsity

This paper is now available on arxiv: http://arxiv.org/abs/1103.0102 In multi-label learning, each sample is associated with several labels. Existing works indicate that exploring correlations between labels improve the prediction performance. However, embedding the label correlations into the training process significantly … Continue reading

Posted in Tianyi's work | Tagged , , , , , , | Leave a comment