I am a Ph.D. student in the Machine Learning Department at CMU.
My advisor is Aarti Singh.

Research

I am currently researching feature selection methods for high dimensional, noisy, and not fully observed clustering problems.

I am also interested in so called safe semi-supervised learning methods, i.e. semi-supervised learning algorithms that take advantage of unlabeled data when possible, but gracefully reduce to supervised learning when unlabeled data are not helpful.

My other areas of interest include general entropy learning theory, and subspace detection and learning under compressive measurment models.

Publications and preprints

Efficient Sparse Clustering of High-Dimensional Non-spherical Gaussian Mixtures
Martin Azizyan, Aarti Singh, Larry Wasserman
arXiv

Feature Selection For High-Dimensional Clustering
Martin Azizyan, Aarti Singh, Larry Wasserman
arXiv

Subspace Learning from Extremely Compressed Measurements
Akshay Krishnamurthy, Martin Azizyan, Aarti Singh
arXiv

RNA design rules from a massive open laboratory
Jeehyung Lee, Wipapat Kladwang, Minjae Lee, Daniel Cantu, Martin Azizyan, Hanjoo Kim, Alex Limpaecher, Sungroh Yoon, Adrien Treuille, Rhiju Das, and EteRNA Participants
PNAS, 2014. (link)

Minimax Theory for High-dimensional Gaussian Mixtures with Sparse Mean Separation
Martin Azizyan, Aarti Singh, Larry Wasserman
NIPS, 2013. (arXiv)

Density-Sensitive Semisupervised Inference
Martin Azizyan, Aarti Singh, Larry Wasserman
Annals of Statistics, 2013. (full text; arXiv version; R code)

Subspace detection of high-dimensional vectors using compressive sampling
Martin Azizyan, Aarti Singh
IEEE Statistical Signal Processing Workshop, 2012. (full text of revised version, PDF)

"Did You See Bob?": Human Localization using Mobile Phones
Ionut Constandache, Xuan Bao, Martin Azizyan, Romit Roy Choudhury
ACM MobiCom, September 2010. (full text)

SurroundSense: Mobile Phone Localization Via Ambience Fingerprinting
Martin Azizyan, Ionut Constandache, Romit Roy Choudhury
ACM MobiCom, September 2009. (full text)