H O M E

I am an assistant professor in the Department of Computer Science at the University of Massachusetts Amherst. I was previously a fellow of both the Pacific Institute for the Mathematical Sciences and the Killam Trusts at the University of British Columbia where I was based in the Laboratory for Computational Intelligence in the Department of Computer Science. I completed my PhD in machine learning in the Department of Computer Science at the University of Toronto.

Research Interests: My research interests lie at the intersection of artificial intelligence, machine learning and statistics. I am particularly interested in hierarchical graphical models and approximate inference/learning techniques including Markov Chain Monte Carlo and variational Bayesian methods. I am also interested in the study of non-likelihood-based inductive principles for statistical models and the trade-off between statistical consistency/efficiency and computational efficiency. I am interested in a broad range of applications for these modeling and learning techniques including classification, collaborative filtering, ranking, unsupervised structure discovery, feature induction, object recognition/image labeling and medical informatics.

Upcoming Papers:

• [Sept 1, 2011] Benjamin M. Marlin, Dave Kale, Robinder Khemani and Randall Wetzel. Unsupervised Pattern Discovery in Electronic Health Care Data using Probabilistic Clustering Models. To appear at the 2nd ACM SIGHIT International Health Informatics Symposium. Januray 28-30, 2012.

Recent Papers:

• [June 1, 2011] Benjamin M. Marlin and Nando de Freitas. Asymptotic Efficiency of Deterministic Estimators for Discrete Energy-Based Models: Ratio Matching and Pseudolikelihood. Proceedings of the The 27th Conference on Uncertainty in Artificial Intelligence.
• [May 3, 2011] Benjamin M. Marlin, Richard S. Zemel, Sam T. Roweis and Malcolm Slaney. Recommender Systems: Missing Data and Statistical Model Estimation. Proceedings of the 22nd International Joint Conference on Artificial Intelligence. (Best papers track).
• [May 3, 2011] Benjamin M. Marlin, Mohammad Emtiyaz Khan, and Kevin Murphy. Piecewise Bounds for Estimating Bernoulli-Logistic Latent Gaussian Models. Proceedings of the 28th International Conference on Machine Learning.
• [May 3, 2011] Kevin Swersky, Marc'Aurelio Ranzato, David Buchman, Benjamin M. Marlin and Nando de Freitas. On Autoencoders and Score Matching for Energy Based Models. Proceedings of the 28th International Conference on Machine Learning.
• [May 3, 2011] David Duvenaud, Benjamin M. Marlin and Kevin Murphy. Multiscale Conditional Random Fields for Semi-supervised Labeling and Classification. Proceedings of the Eighth Canadian Conference on Computer and Robot Vision, 2011.

PhD Thesis: My PhD thesis is titled Missing Data Problems in Machine Learning. It deals with the problem of unsupervised learning in the presence of non-random missing data, as well as the problem of classification in the presence of missing features. The work on non-random missing data is motivated by the problem of rating prediction in collaborative filtering, and uses a new data set collected at Yahoo! Research with Malcolm Slaney. The work on classification with missing features focuses on medical decision making using standard data sets, as well as higher dimensional tasks like digit classification with missing pixels. [Thesis Abstract] [Thesis PDF] [Short Defense Slides PDF] [Long Defense Slides PDF]

Contact Information:
marlin@cs.umass.edu
140 Governors Drive, Office 234
Amherst, MA 01003

This site last updated September 12, 2011. © Benjamin M. Marlin