Daniel Hsu

I am an assistant professor in the Department of Computer Science at Columbia University in the City of New York.
I am also a member of the Institute for Data Sciences and Engineering.

My research interests are in algorithmic statistics, machine learning, and privacy.


Coordinates
Office
702 CEPSR
E-mail
<username>@cs.<domain>
Post
500 W. 120th St., M.C. 0401
New York, NY 10027-7003
Phone
555-555-5555

Teaching

Fall 2014 COMS 4772 Advanced Machine Learning
Spring 2015 COMS 4771 Elementary Machine Learning
Past courses

Research papers

Preprints

Loss minimization and parameter estimation with heavy tails.
Daniel Hsu and Sivan Sabato.
Preprint, 2013.
[arxiv, slides for a related talk]

Analysis of a randomized approximation scheme for matrix multiplication.
Daniel Hsu, Sham M. Kakade, and Tong Zhang.
Preprint, 2012.
[arxiv]


Postprints

Tensor decompositions for learning latent variable models.
Anima Anandkumar, Rong Ge, Daniel Hsu, Sham M. Kakade, and Matus Telgarsky.
Journal of Machine Learning Research, 15(Aug):2773-2832, 2014.
[pdf, link, slides for a related talk]

Scalable nonlinear learning with adaptive polynomial expansions.
Alekh Agarwal, Alina Beygelzimer, Daniel Hsu, John Langford, and Matus Telgarsky.
Advances in Neural Information Processing Systems 27, 2014.
[pdf, arxiv]

The large margin mechanism for differentially private maximization.
Kamalika Chaudhuri, Daniel Hsu, and Shuang Song.
Advances in Neural Information Processing Systems 27, 2014.
[pdf, arxiv]

A spectral algorithm for learning class-based n-gram models of natural language.
Karl Stratos, Do-kyum Kim, Michael Collins, and Daniel Hsu.
Thirtieth Conference on Uncertainty in Artificial Intelligence, 2014.
[pdf, code by Karl, more code by Karl]

Taming the monster: a fast and simple algorithm for contextual bandits.
Alekh Agarwal, Daniel Hsu, Satyen Kale, John Langford, Lihong Li, and Robert E. Schapire.
Thirty-First International Conference on Machine Learning, 2014.
[pdf, arxiv, link, slides]

Heavy-tailed regression with a generalized median-of-means.
Daniel Hsu and Sivan Sabato.
Thirty-First International Conference on Machine Learning, 2014.
[arxiv, link]

A tensor approach to learning mixed membership community models.
Anima Anandkumar, Rong Ge, Daniel Hsu, and Sham M. Kakade.
Journal of Machine Learning Research, 15(Jun):2239-2312, 2014.
[pdf, arxiv, link]

Random design analysis of ridge regression.
Daniel Hsu, Sham M. Kakade, and Tong Zhang.
Foundations of Computational Mathematics, 14(3):569-600, 2014.
[pdf, arxiv, link]

Contrastive learning using spectral methods.
James Zou, Daniel Hsu, David Parkes, and Ryan Adams.
Advances in Neural Information Processing Systems 26, 2013.
[pdf]

When are overcomplete topic models identifiable? Uniqueness of tensor Tucker decompositions with structured sparsity.
Anima Anandkumar, Daniel Hsu, Majid Janzamin, Sham M. Kakade.
Advances in Neural Information Processing Systems 26, 2013.

Stochastic convex optimization with bandit feedback.
Alekh Agarwal, Dean P. Foster, Daniel Hsu, Sham M. Kakade, and Alexander Rakhlin.
SIAM Journal on Optimization, 23(1):213-240, 2013.
[pdf, arxiv, link]

A tensor spectral approach to learning mixed membership community models.
Anima Anandkumar, Rong Ge, Daniel Hsu, and Sham M. Kakade.
Twenty-Sixth Annual Conference on Learning Theory, 2013.
[pdf (short version), arxiv (full version)]

Learning linear Bayesian networks with latent variables.
Anima Anandkumar, Daniel Hsu, Adel Javanmard, and Sham M. Kakade.
Thirtieth International Conference on Machine Learning, 2013.
[pdf (short version), arxiv (full version)]

Learning mixtures of spherical Gaussians: moment methods and spectral decompositions.
Daniel Hsu and Sham M. Kakade.
Fourth Innovations in Theoretical Computer Science, 2013.
[pdf (short version), arxiv (full version), talk slides]

A spectral algorithm for learning hidden Markov models.
Daniel Hsu, Sham M. Kakade, and Tong Zhang.
Journal of Computer and System Sciences, 78(5):1460-1480, 2012.
[pdf, arxiv, link, errata]

A tail inequality for quadratic forms of subgaussian random vectors.
Daniel Hsu, Sham M. Kakade, and Tong Zhang.
Electronic Communications in Probability, 17(52):1-6, 2012.
[pdf, link]

A spectral algorithm for latent Dirichlet allocation.
Anima Anandkumar, Dean P. Foster, Daniel Hsu, Sham M. Kakade, and Yi-Kai Liu.
Advances in Neural Information Processing Systems 25, 2012.
[pdf (short version), arxiv (full version)]

Identifiability and unmixing of latent parse trees.
Daniel Hsu, Sham M. Kakade, and Percy Liang.
Advances in Neural Information Processing Systems 25, 2012.
[pdf, arxiv]

Learning mixtures of tree graphical models.
Anima Anandkumar, Daniel Hsu, Furong Huang, and Sham M. Kakade.
Advances in Neural Information Processing Systems 25, 2012.

Convergence rates for differentially private statistical estimation.
Kamalika Chaudhuri and Daniel Hsu.
Twenty-Ninth International Conference on Machine Learning, 2012.
[pdf]

A method of moments for mixture models and hidden Markov models.
Anima Anandkumar, Daniel Hsu, and Sham M. Kakade.
Twenty-Fifth Annual Conference on Learning Theory, 2012.
[pdf, arxiv, talk slides, slides for a related talk]

Random design analysis of ridge regression.
Daniel Hsu, Sham M. Kakade, and Tong Zhang.
Twenty-Fifth Annual Conference on Learning Theory, 2012.
[pdf (short version), arxiv (full version)]

Tail inequalities for sums of random matrices that depend on the intrinsic dimension.
Daniel Hsu, Sham M. Kakade, and Tong Zhang.
Electronic Communications in Probability, 17(14):1-13, 2012.
[pdf, link, errata]

Spectral methods for learning multivariate latent tree structure.
Anima Anandkumar, Kamalika Chaudhuri, Daniel Hsu, Sham M. Kakade, Le Song, and Tong Zhang.
Advances in Neural Information Processing Systems 24, 2011.
[pdf, arxiv]

Stochastic convex optimization with bandit feedback.
Alekh Agarwal, Dean P. Foster, Daniel Hsu, Sham M. Kakade, and Alexander Rakhlin.
Advances in Neural Information Processing Systems 24, 2011.
[pdf (journal version), arxiv]

Efficient optimal learning for contextual bandits.
Miroslav Dudik, Daniel Hsu, Satyen Kale, Nikos Karampatziakis, John Langford, Lev Reyzin, and Tong Zhang.
Twenty-Seventh Conference on Uncertainty in Artificial Intelligence, 2011.
[pdf]

Sample complexity bounds for differentially private learning.
Kamalika Chaudhuri and Daniel Hsu.
Twenty-Fourth Annual Conference on Learning Theory, 2011.
[pdf, talk slides]

Robust matrix decomposition with sparse corruptions.
Daniel Hsu, Sham M. Kakade, and Tong Zhang.
IEEE Transactions on Information Theory, 57(11):7221-7234, 2011.
[pdf, link]

Agnostic active learning without constraints.
Alina Beygelzimer, Daniel Hsu, John Langford, and Tong Zhang.
Advances in Neural Information Processing Systems 23, 2010.
[pdf]

An online learning-based framework for tracking.
Kamalika Chaudhuri, Yoav Freund, and Daniel Hsu.
Twenty-Sixth Conference on Uncertainty in Artificial Intelligence, 2010.
[pdf]

Algorithms for active learning.
Daniel Hsu
Ph.D. dissertation, University of California, San Diego, 2010.
[pdf]

Multi-label prediction via compressed sensing.
Daniel Hsu, Sham M. Kakade, John Langford, and Tong Zhang.
Advances in Neural Information Processing Systems 22, 2009.
[pdf, arxiv, talk slides]

A parameter-free hedging algorithm.
Kamalika Chaudhuri, Yoav Freund, and Daniel Hsu.
Advances in Neural Information Processing Systems 22, 2009.
[pdf, note about epsilon-quantile regret]

A spectral algorithm for learning hidden Markov models.
Daniel Hsu, Sham M. Kakade, and Tong Zhang.
Twenty-Second Annual Conference on Learning Theory, 2009.
[pdf (journal version), arxiv, errata]

Hierarchical sampling for active learning.
Sanjoy Dasgupta and Daniel Hsu.
Twenty-Fifth International Conference on Machine Learning, 2008.
[pdf, code]

A general agnostic active learning algorithm.
Sanjoy Dasgupta, Daniel Hsu, and Claire Monteleoni.
Advances in Neural Information Processing Systems 20, 2007.
[pdf]

On-line estimation with the multivariate Gaussian distribution.
Sanjoy Dasgupta and Daniel Hsu.
Twentieth Annual Conference on Learning Theory, 2007.
[pdf]

A concentration theorem for projections.
Sanjoy Dasgupta, Daniel Hsu, and Nakul Verma.
Twenty-Second Conference on Uncertainty in Artificial Intelligence, 2006.
[pdf]

Service

Senior PCs
Conference on Learning Theory (COLT) 2011, 2013
International Conference on Machine Learning (ICML) 2012, 2013
Conference on Neural Information Processing Systems (NIPS) 2012, 2013
Program Committees
AAAI Conference on Artificial Intelligence (AAAI)
International Conference on Artificial Intelligence and Statistics (AISTATS)
International Conference on Machine Learning (ICML)
Conference on Neural Information Processing Systems (NIPS)
Conference on Uncertainty in Artificial Intelligence (UAI)
Tutorials
AAAI 2014 Tensor decompositions for learning latent variable models (July 28, 2014)
ICML 2013 Tensor decomposition methods for latent variable model estimation (June 16, 2013)
Workshops
ICML 2014 Method of moments and spectral learning (June 25, 2014)
DIMACS/CCICADA Systems and analytics of big data (March 17–18, 2014)
NIPS 2013 Spectral learning (December 10, 2013)
ICML 2013 Spectral learning (June 21, 2013)

&c.

Information
Information for prospective students
Biosketch
Groups
Machine learning
Theory of computation
Foundations of data science
Software
BOINC (software for volunteer computing)
Vowpal Wabbit (fast online learning software)
Other stuff
Endorsements
Perl script for producing Bkoos
MATLAB codes for estimation via method-of-moments:
Shell script to convert PDF slides to Keynote format
Website design ripped off from Brian McFee.