INVITED AND TUTORIAL SPEAKERSJames Cussens (University of York)
Jason Eisner (Johns Hopkins University)
Jure Leskovec (Cornell University)
Raymond Mooney (University of Texas at Austin)
Scott Sanner (NICTA)
Philip S. Yu (University of Illinois at Chicago)
James CussensA tutorial on logic-based approaches to SRL
The relations in Statistical Relational Learning are often expressed using first-order logic, leading to formalisms which combine both logical and probabilistic representations. In this talk I intend to explain the most important consequences of adopting a logical approach to SRL. Defining distributions over 'possible worlds' is a common theme to many such approaches. Two prominent logic-based formalisms - Markov logic networks and PRISM programs - will be used as exemplars. Although the talk is tutorial in nature, I hope to make it interesting to those already familiar with this area!
Jason EisnerWeighted Deduction as an Abstraction Level for AI
The field of AI has become implementation-bound. We have plenty of ideas, but it is increasingly laborious to try them out, as our models become more ambitious and our datasets become larger, noisier, and more heterogeneous. The software engineering burden makes it hard to start new work; hard to reuse and combine existing ideas; and hard to educate our students. In this talk, I'll propose to hide many common implementation details behind a new level of abstraction that we are developing. Dyna is a declarative programming language that combines logic programming with functional programming. It also supports modularity. It may be regarded as a kind of deductive database, theorem prover, truth maintenance system, or equation solver. I will illustrate how Dyna makes it easy to specify the combinatorial structure of typical computations needed in natural language processing, machine learning, and elsewhere in AI. Then I will sketch implementation strategies and program transformations that can help to make these computations fast and memory-efficient. Finally, I will suggest that machine learning should be used to search for the right strategies for a program on a particular workload.
Jason Eisner is Associate Professor of Computer Science at Johns Hopkins University, where he is also affiliated with the Center for Language and Speech Processing, the Cognitive Science Department, and the national Center of Excellence in Human Language Technology. He is particularly interested in designing algorithms that statistically exploit linguistic structure. His 60+ papers have presented a number of algorithms for parsing and machine translation; algorithms for constructing and training weighted finite-state machines; formalizations, algorithms, theorems and empirical results in computational phonology; and unsupervised or semi-supervised learning methods for domains such as syntax, morphology, and word-sense disambiguation.
Raymond MooneyBottom-up Search and Transfer Learning in SRL
This talk addresses two important issues motivated by of our recent research in SRL. First, is the value of data-driven, "bottom-up" search in learning the structure of SRL models. Bottom-up induction has a long history in traditional ILP; however, its use in SRL has been somewhat limited. We review recent results on several structure-learning methods for Markov Logic Networks (MLNs) that highlight the value of bottom-up search. Second, is the value of transfer learning in reducing the data and computational demands of SRL. By inducing a predicate mapping between seemingly disparate domains, effective SRL models can be efficiently learned from very small amounts of in-domain training data. For example, by transferring a model learned from data about a CS department, we have induced reasonably accurate models for IMDB movie data given training data about only a single person.
Raymond J. Mooney is a Professor in the Department of Computer Sciences at the University of Texas at Austin. He received his Ph.D. in 1988 from the University of Illinois at Urbana/Champaign. He is an author of over 150 published research papers, primarily in the areas of machine learning and natural language processing. He is the current President of the International Machine Learning Society, was program co-chair for the 2006 AAAI Conference on Artificial Intelligence, general chair of the 2005 Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing, and co-chair of the 1990 International Conference on Machine Learning. He is a Fellow of the American Association for Artificial Intelligence and recipient of best paper awards from the National Conference on Artificial Intelligence, the SIGKDD International Conference on Knowledge Discovery and Data Mining, the International Conference on Machine Learning, and the Annual Meeting of the Association for Computational Linguistics. His recent research has focused on learning for natural-language processing, connecting language and perception, statistical relational learning, and transfer learning.
Scott SannerFirst-order Models for Sequential Decision-making: Insights, Caveats, and Tricks-of-the-Trade
In this talk I will discuss first-order models and algorithms for sequential decision-making, specifically those approaches that admit exact lifted solutions. The first emphasis of the talk will be on the insights that underlie these models and algorithms along with potential caveats for their practical application. The second emphasis of the talk will be on a variety of extensions of the first-order Markov decision process (MDP) framework such as the factored first-order MDP and the first-order partially observable MDP. The third emphasis of the talk will be on the algorithmic tricks-of-the-trade that allow the practical application of these models; this includes (a) useful data structures, (b) efficient solution techniques for first-order linear programs, (c) new techniques for first-order variable elimination, and (d) practical methods for maintaining compact, consistent first-order representations without theorem proving.
Philip S. YuScalable Link Mining and Analysis on Information Networks
With the ubiquity of information networks and their broad applications, there have been numerous studies on the construction, online analytical processing, and mining of information networks in multiple disciplines, including social network analysis, World-Wide Web, database systems, data mining, machine learning, and networked communication and information systems. Algorithms like PageRank and HITS have been developed in late 1990s to explore links among Web pages to discover authoritative pages and hubs. Links have also been popularly used in citation analysis and social network analysis. However, there is a lack of systematic treatment on how to fully explore the power of links in scalable data analysis. In this talk, the power of links are examined in details to improve the effectiveness and efficiency of typical data analysis tasks, including information integration, on-line analytic processing, and other interesting data mining tasks, especially in the multi-relational databases and/or the World-Wid e Web environments.
Philip S. Yu is a Professor in the Department of Computer Science at the University of Illinois at Chicago and also holds the Wexler Chair in Information Technology. He was manager of the Software Tools and Techniques group at the IBM Thomas J. Watson Research Center. Dr. Yu is a Fellow of the ACM and the IEEE. He served as the Editor-in-Chief of IEEE Transactions on Knowledge and Data Engineering (2001-2004). He is an associate editor of ACM Transactions on Knowledge Discovery from Data and also ACM Transactions of the Internet Technology. He serves on the steering committee of IEEE Int. Conference on Data Mining. He was a member of the IEEE Data Engineering steering committee. Dr. Yu received a Research Contributions Award from IEEE Intl. Conference on Data Mining in 2003. His research interests include data mining, and database systems. He has published more than 540 papers in refereed journals and conferences. He holds or has applied for more than 300 US patents. Dr. Yu was an IBM Master Inventor.