September 2011 - talks - lists.cs.princeton.edu

Colloquium Speaker Eric Xing: Thrus, Oct 6th- 4:30pm
by Nicole E. Wagenblast 04 Oct '11

04 Oct '11

Jointly Maximum Margin and Maximum Entropy Learning of Graphical Models Eric Xing, Carnegie Mellon University Thursday, October 6, 2011- 4:30pm Computer Science, 105 Graphical models (GMs) offer a powerful language to elegantly define expressive distributions, and a generic computational framework to support reasoning under uncertainty in a wide range of problems. Popular paradigms for training GMs include the maximum likelihood estimation, and more recently the max-margin learning, each enjoys some advantages, as well as weaknesses. For example, the maximum margin structured prediction model such as M3N lacks a straightforward probabilistic interpretation of the learning scheme and the prediction rule. Therefore its unique advantages such as support vector sparsity and kernel tricks cannot be easily conjoined with the merits of a probabilistic model such as Bayesian regularization, model averaging, and ability to model hidden variables. In this talk, I present a new general framework called Maximum Entropy Discrimination Markov Networks (MEDN), which integrates the margin-based and likelihood-based approaches and combines and extends their merits. This new learning paradigm naturally facilitates integration of the generative and discriminative principles under a unified framework, and the basic strategies can be generalized to learn arbitrary GMs, such as the generative Bayesian networks, models with structured hidden variables, and even nonparametric Bayesian models, with a desirable maximum margin effect on structured or unstructured predictions. I will discuss a number of theoretical properties of this approach, and show applications of MEDN to learning a wide range of GMs including: fully supervised structured i/o model, max-margin structured i/o models with hidden variables, a max-margin LDA-style model for jointly discovering “discriminative” latent topics and predicting document label/score of text documents, or total scene and objective categories in natural images, etc. Our empirical results strongly suggest that, for any GM with structured or unstructured labels, MEDN always leads to a more accurate predictive GM than the one trained under either MLE or Max Margin. Joint work with Jun Zhu. Dr. Eric Xing is an associate professor in the School of Computer Science at Carnegie Mellon University. His principal research interests lie in the development of machine learning and statistical methodology; especially for solving problems involving automated learning, reasoning, and decision-making in high-dimensional and dynamic possible worlds; and for building quantitative models and predictive understandings of biological systems. Professor Xing received a Ph.D. in Molecular Biology from Rutgers University, and another Ph.D. in Computer Science from UC Berkeley. His current work involves, 1) foundations of statistical learning, including theory and algorithms for estimating time/space varying-coefficient models, sparse structured input/output models, and nonparametric Bayesian models; 2) computational and statistical analysis of gene regulation, genetic variation, and disease associations; and 3) application of statistical learning in social networks, computer vision, and natural language processing. Professor Xing has published over 140 peer-reviewed papers, and is an associate editor of the Annals of Applied Statistics, the PLoS Journal of Computational Biology, and an Action Editor of the Machine Learning journal. He is a recipient of the NSF Career Award, the Alfred P. Sloan Research Fellowship in Computer Science, and the United States Air Force Young Investigator Award.

1 1

Colloquium speaker Swarat Chaudhuri Thurs Sept 29
by Nicole E. Wagenblast 28 Sep '11

28 Sep '11

Composing Composure: Reasoning about Robustness of Software Systems Swarat Chaudhuri, Rice University Thursday, September 29- 4:30pm Computer Science 105 Most software is inherently nonrobust--change the operating conditions of a typical program slightly, and you may obtain very different results. This lack of robustness can not only result in unpredictable runtime behavior, it also makes testing, approximation, and mathematical analysis of programs highly challenging. We will argue that methods for automated logical reasoning about programs provide a way to cope with this problem. Using a program analysis, we can sometimes determine if a given program is robust, and if so, exploit this property. On the other hand, if a program is not robust, logic can help us approximate it into one that is robust. Using applications from several different areas of computer science, I will show how these techniques can lead to more reliable program execution under uncertainty, opportunities for language-based approximate computation, and easier solutions to hard optimization problems involving programs. Swarat Chaudhuri is an Assistant Professor of Computer Science at Rice University. His research lies in the interface of automated reasoning and programming languages.

1 1

Distinguished Colloquium speaker, Cynthia Dwork, Wed Sept 21- 4:30pm
by Nicole E. Wagenblast 15 Sep '11

15 Sep '11

Distinguished Colloquium Series Speaker Fairness Through Awareness Cynthia Dwork, Microsoft Research, SVC Wednesday, September 21, 2011, 4:30 PM Computer Science 105 "Why was I not shown this advertisement? Why was my loan application denied? Why was I denied admission to this university?" In this work we initiate the formal study of fairness in classification, where the goal is to prevent discrimination against protected population subgroups in classification systems while simultaneously preserving utility for the party carrying out the classification (eg, the advertiser, bank, or admissions committee). We argue that a classification is fair only when individuals who are similar with respect to the classification task at hand are treated similarly, and this in turn requires understanding of sub cultures of the population. Similarity metrics are applied in many contexts, but these are often hidden. Our work explicitly exposes the metric, opening it to public debate. We then formalize and show how to achieve fairness in classification, given a similarity metric. We also give conditions on the metric under which our "local" notion ensures statistical parity: namely that the demographics of those receiving any given classification are the same as the demographics of the underlying population. In a complementary setting, we propose tools for what can be viewed as "fair affirmative action." Namely, we give methods for guaranteeing statistical parity for a group while treating similar individuals as similarly as possible. Finally, we discuss the relationship of fairness to privacy: to what extent might fairness imply privacy, and is it possible to employ tools developed in the context of differential privacy in order to obtain fairness. Joint work with Moritz Hardt, Toniann Pitassi, Omer Reingold, and Richard Zemel. Cynthia Dwork '79, Distinguished Scientist at Microsoft Research, is the world's foremost expert on placing privacy-preserving data analysis on a mathematically rigorous foundation. A cornerstone of this work is differential privacy, a strong privacy guarantee permitting highly accurate data analysis. Dr. Dwork has also made seminal contributions in cryptography and distributed computing, and is a recipient of the Edsger W. Dijkstra Prize, recognizing some of her earliest work establishing the pillars on which every fault-tolerant system has been built for decades. She is a member of the US National Academy of Engineering and a Fellow of the American Academy of Arts and Sciences.

1 0

Wei Dong FPO
by Melissa M. Lawson 08 Sep '11

08 Sep '11

Wei Dong will present his FPO on Wednesday September 14 at 9:30AM in Room 402. The members of his committee are: Kai Li, advisor; Moses Charikar and Andrea LaPaugh, readers; Brian Kernighan and Olga Troyanskaya, nonreaders. A copy of his thesis, "High-Dimensional Similarity Search for Large Datasets", is on display in Room 310. Everyone is invited to attend his talk.

1 0