WebEvent: Search for BISC

Tuesday, September 14, 2004

	Sep 14	4 p.m. - 5:30 p.m.	Web Intelligence, World Knowledge and Fuzzy Logic Location: 405 Soda Hall Seminar Speaker Name: L. A. Zadeh Seminar Speaker Affil.: BISC Program; EECS, UC Berkeley Seminar Series: BISC seminar For more information: http://www-bisc.cs.berkeley.edu Details: Existing search engines—with Google at the top—have many remarkable capabilities; but what is not among them is deduction capability—the capability to synthesize an answer to a query from bodies of information which reside in various parts of the knowledge base. In recent years, impressive progress has been made in enhancing performance of search engines through the use of methods based on bivalent logic and bivalent-logic-based probability theory. But can such methods be used to add nontrivial deduction capability to search engines, that is, to upgrade search engines to question-answering systems? A view which is articulated in this note is that the answer is “No.” The problem is rooted in the nature of world knowledge, the kind of knowledge that humans acquire through experience and education. It is widely recognized that world knowledge plays an essential role in assessment of relevance, summarization, search and deduction. But a basic issue which is not addressed is that much of world knowledge is perception-based, e.g., “it is hard to find parking in Paris,” “most professors are not rich,” and “it is unlikely to rain in midsummer in San Francisco.” The problem is that (a) perception-based information is intrinsically fuzzy; and (b) bivalent logic is intrinsically unsuited to deal with fuzziness and partial truth. To come to grips with the fuzziness of world knowledge, new tools are needed. The principal new tool—a tool which is briefly described in their note—is Precisiated Natural Language (PNL). PNL is based on fuzzy logic and has the capability to deal with partiality of certainty, partiality of possibility and partiality of truth. These are the capabilities that are needed to be able to draw on world knowledge for assessment of relevance, and for summarization, search and deduction.
Tuesday, September 21, 2004

	Sep 21	4 p.m. - 5:30 p.m.	Text Summarization -- In Search of Effective Ideas and Techniques Location: 405 Soda Hall Seminar Speaker Name: Shuhua Liu Seminar Speaker Affil.: Åbo Akademi University, Finland & and BISC Program; EECS, UC Berkeley Seminar Series: BISC seminar For more information: http://www-bisc.cs.berkeley.edu Details: The dominant approach to text summarization is selection-based, by which the most content-bearing sentences or passages are identified and selected to compose a summary. However, the results from such a process often suffer from flaws such as incoherent content and poor readability due to unclear relationships between the selected text excerpts, dangling references, and so on. In our work we follow a process model of text summarization based on topic analysis. By directing the content selection process with identified topic structure of the text, the logical relation between the selected sentences is expected to be captured and better presented, which will also be valuable resource for the understanding of text meaning. Key issues and relevant techniques for the implementation of the model will be discussed.
Tuesday, September 28, 2004

	Sep 28	4 p.m. - 5:30 p.m.	Decision Tree using Evolutionary Techniques: GA-GP-Based Fuzzy Decison Tree Model Location: 405 Soda Hall Seminar Speaker Name: Souad Souafi Bensafi Seminar Speaker Affil.: BISC Program, EECS Department, UC Berkeley Seminar Series: BISC For more information: http://www-bisc.cs.berkeley.edu Details: Our aim is to develop intelligent computing techniques that address the problem of multi-criteria decision making dealing with subjective and imprecise data. This kind of problem requires conception of intelligent systems able to replace a human with expertise in a specific domain in a decision making process. So, the intelligent system should take into account the subjective and imprecise character of data on the one hand, and represent the user or expert's preferences and knowledge on the other hand. For this purpose, we developed a generic multi-criteria model based on fuzzy logic concepts for decision support systems. Our goal is to build such a model by (1) fitting real-world data and (2) representing the preferences of specific-domain users or experts. Toward this end, we used evolutionary computation techniques. Initially, we worked on a first order aggregation model and performed its learning using genetic algorithms, in which these preferences have been represented by a weighting vector associated with the variables involved in the aggregation process. This has been used in a specific application related to university admissions. Then, we developed a more advanced multi-aggregation model based on a hierarchical decision trees and for the learning process of this model, we developed a technique inspired from genetic programming. In this model tree nodes represent aggregators, terminals or leaves correspond to variables, and weight values are added to the children branches for each aggregator. The aggregation result overall the variables is then obtained by running recursively the root aggregator of the tree.
Tuesday, October 5, 2004

	Oct 5	4 p.m. - 5:30 p.m.	The Failure of Clustering in Search User Interfaces Location: 405 Soda Hall Seminar Speaker Name: Marti Hearst Seminar Speaker Affil.: SIMS, UC Berkeley Seminar Series: BISC For more information: http://www-bisc.cs.berkeley.edu Details: Time and again, researchers (including the speaker, long ago) have proposed using text or image clustering for search user interfaces, despite that fact that usability studies consistently show negative results for general search (as opposed to for analysis). In this talk I will summarize the evidence against using clustering in general search user interfaces and describe why it fails. To avoid this being an entirely negative talk, I will point the way towards what appears to be the solution, which is the flexible and intuitive presentation of categories in search user interfaces. Speaker Bio: Dr. Marti Hearst is an associate professor in SIMS, with an affiliate appointment in the CS Division at UCB. Her primary research interests are user interfaces and visualization for information retrieval, empirical computational linguistics, and text data mining. She received BA, MS, and PhD degrees in Computer Science from the University of California at Berkeley, and she was a Member of the Research Staff at Xerox PARC from 1994 to 1997. Prof. Hearst is on the editorial boards of ACM Transactions on Information Systems and ACM Transactions on Computer-Human Interaction and was formerly on the boards of Computational Linguistics and IEEE Intelligent Systems, and was the program co-chair of HLT-NAACL '03 and SIGIR '99. She has received an NSF CAREER award, an IBM Faculty Award, an Okawa Foundation Fellowship, and two student-initiated Excellence in Teaching awards.
Tuesday; Oct 26, 2004
	Oct 26	Tuesday; 1:30-2:30pm	Semantics – the implicit, the formal and the powerful; (with a case study in Glycomics) Location: 606 Soda Hall Seminar Speaker Name: Amit Sheth Seminar Speaker Affil.: Large Scale Distributed Information Systems (LSDIS) lab, Univ. of Georgia Seminar Series: BISC For more information: http://www-bisc.cs.berkeley.edu Details: Semantics has been recognized as the key to next generation of more powerful information systems for better search, integration, question/answering as well as analysis/discovery. Semantics has long been studied in many disciplines including linguistics, AI, IR, information and database systems, and soft computing, and a rich variety of approaches, techniques and tools have been developed. More recently, the Semantic Web community has made concerted effort in using semantics by defining standards for the modeling of knowledge based on Description Logic (DL) based languages such as OWL, and focused on corresponding reasoning techniques that have rather narrow set of applications. We view these recent approaches as addressing a subset of challenges, complemented by techniques that deal with broad variety of information and knowledge, expressiveness, computational capabilities and computability. In this talk we attempt to organize this broad variety of options from a pedagogic perspective that characterizes semantic approaches as implicit (such as those based on statistical and machine learning), formal (such as those based on DL to FOL) and powerful (such as those based on soft computing). To exemplify this perspective, we look at some examples from the domain of life sciences which offer rich sets of challenges due to the complexity of biological systems. More specifically we look at our current research in Glycomics that involve the requirements for creation of taxonomies and ontologies with higher expressive representations, semantic annotation of textual data in heterogeneous formats as well as machine generated scientific data, semantic search of scientific literature, semantic integration of heterogeneous textual and scientific (nontextual) data, wrapping of data analysis tools with semantically annotated Web Services, and development of semantic web processes leading to better and quicker interpretation/analytics and discovery. In particular, we will offer concrete examples demonstrating the need for more expressive representation that follow the ideas offered by Prof. Zadeh in “Toward a perception-based theory of probabilistic reasoning with imprecise probabilities”. Among the novel research outcomes including automatic taxonomy generation in Taxaminer, development of a comprehensive ontology for Glycomics called GlycO, early efforts in semantic annotation of machine generated scientific data, and preliminary ideas about fpOWL, an extension to the ontology language OWL that allows probabilistic and fuzzy reasoning. See project at the LSDIS lab for more information. Acknowledgement: Will York, Christopher Thomas, and other members of Bioinformatics for Glycan Express project team; Cartic Ramakrishnan and members of Taxaminer team. About the speaker: Amit Sheth is an Educator, Researcher and Entrepreneur. He joined the University of Georgia and started the LSDIS lab in 1994. Earlier, he served in R&D groups at Bellcore (now Telcordia Technologies), Unisys, and Honeywell. In August 1999, Sheth founded Taalee, Inc., a VC funded enterprise software and internet infrastructure startup based on the technology developed at the LSDIS lab. He managed Taalee as its CEO until June 2001. Following Taalee's acquisition/merger, he serves as the CTO and co-founder of Semagix, Inc. (formerly Voquette, Inc). His research has led to several commercial products and applications. He has published over 175 papers and articles (in the areas of semantic interoperability, federated databases, workflow management, Semantic Web), given over 130 invited talks and colloquia including 19 keynotes, (co)-organized/chaired twelve conferences/workshops, and served on over 90 program committees. He is a member of W3C Advisory Committee, SWSA, etc. http://lsdis.cs.uga.edu/~amit and http://www.semagix.com/company_team.html
Tuesday, Nov 9, 2004
	Nov 9	4:00-5:30pm	Web Search as a Computational Challenge Location: 306 Soda Hall Seminar Speaker Name: Peter Norvig Seminar Speaker Affil.: Google Inc Seminar Series: BISC Seminar For more information: http://www-bisc.cs.berkeley.edu Details: For users of the Internet, web search has emerged as the second most popular application, after email. For computer scientists, web search offers challenges in software infrastructure, distributed systems, information retrieval, machine learning, and natural language understanding. This talk will examine these challenges. Short Bio: =========== Peter Norvig is the Director of Search Quality at Google Inc.. He is a Fellow and Councilor of the American Association for Artificial Intelligence and co-author of Artificial Intelligence: A Modern Approach, the leading textbook in the field. Previously he was head of the Computational Sciences Division at NASA Ames Research Center, where he oversaw a staff of 200 scientists performing NASA's research and development in autonomy and robotics, automated software engineering and data analysis, neuro-engineering, collaborative systems research, and simulation-based decision-making. Before that he was Chief Scientist at Junglee, where he helped develop one of the first Internet comparison shopping service; Chief designer at Harlequin Inc; and Senior Scientist at Sun Microsystems Laboratories. Dr. Norvig received a B.S. in Applied Mathematics from Brown University and a Ph.D. in Computer Science from the University of California at Berkeley. He has been a Professor at the University of Southern California and a Research Faculty Member at Berkeley. He has over fifty publications in various areas of Computer Science, concentrating on Artificial Intelligence, Natural Language Processing and Software Engineering including the books Paradigms of AI Programming: Case Studies in Common Lisp, Verbmobil: A Translation System for Face-to-Face Dialog, and Intelligent Help Systems for UNIX.

Wednesday, November 10, 2004
	Nov 10	1:00-2:30pm	Social network analysis of text Location: 380 Soda Hall Seminar Speaker Name: Dragomir R. Radev Seminar Speaker Affil.: University of Michigan Seminar Series: BISC seminar For more information: http://www-bisc.cs.berkeley.edu Details: Textual data is everywhere, in email and scientific papers, in online newspapers and e-commerce sites. The Web contains more than 200 terabytes of text not even counting the contents of dynamic textual databases. This enormous source of knowledge is seriously underexploited. Textual documents on the Web are very hard to model computationally: they are unstructured, time-dependent, collectively authored, and of uneven importance. Traditional grammar-based techniques don't scale up to address such problems. Novel representations and analytical tools are needed. NewsInEssence (www.newsinessence.com) is a system that crawls the Web for news, automatically clusters them by topic, and produces user-defined extractive summaries of each cluster. A recent addition to the battery of summarization algorithms available to NewsInEssence is the Cosine Centrality method. In this talk I will describe how one can apply the theory of social networks and stochastic processes (in particular rank-based prestige and random walks on undirected graphs) to multi-document text summarization. (I will begin my talk with a short tutorial on the mathematics needed for the rest of the talk.) If time permits, at the end of the talk, I will quickly describe two recent ongoing projects in my research group: one on machine learning for object classification using random walks on bipartite (feature-object) graphs and another on using phylogenetic techniques for fact tracking in evolving multi-document summarization. -------------------------------------------- Short Bio: ========== Dragomir R. Radev is Assistant Professor of Information, Electrical Engineering and Computer Science, and Linguistics at the University of Michigan, Ann Arbor. He leads the CLAIR (Computational Lingusitics And Information Retrieval) group which currently includes 12 undergraduate and graduate students. Dragomir holds a Ph.D. in Computer Science from Columbia University. Before joining Michigan, he was a Research Staff Member at IBM's TJ Watson Research Center in Hawthorne, NY. He is the author of more than 45 papers on information retrieval, text summarization, graph models of the Web, question answering, machine translation, text generation, and information extraction. Dr. Radev's current research on probabilistic and link-based methods for exploiting very large textual repositories, representing and acquiring knowledge of genome regulation, and semantic entity and relation extraction from Web-scale text document collections is supported by NSF and NIH. Dragomir serves on the HLT-NAACL advisory committee, was recently reelected as treasurer of NAACL, is a member of the editorial boards of JAIR and Information Retrieval, and is a four-time finalist at the ACM international programming finals (as contestant in 1993 and as coach in 1995-1997). Dragomir received a graduate teaching award at Columbia and recently, the U. of Michigan award for Outstanding Research Mentorship (UROP).
Tuesday, Nov 30, 2004; 405 Soda Hall, 4:00-5:30pm
	Tuesday, Nov 30	4:00-5:30pm	Title: Uncertainty in an unknown world Prof. Stuart Russell Computer Science Division, University of Description: Recent advances in knowledge representation for probability models have allowed for uncertainty about the properties of objects and the relations that might hold among them. Such models, however, typically assume exact knowledge of which objects exist and of which object is which---that is, they assume domain closure and unique names. These assumptions greatly simplify the sample space for probability models, but are inappropriate for many real-world situations. This talk presents a formal language, BLOG, for defining probability models over worlds with unknown objects and in which several terms may refer to the same object. The language has a simple syntax based on first-order logic, combined with local probability functions for quantifying conditional dependencies. A key additional element is the number statement, which specifies a conditional distribution over the number of objects that satisfy a given property. Subject to certain acyclicity constraints, every BLOG model specifies a unique probability distribution over the full set of possible worlds for the first-order language. Furthermore, complete inference algorithms exist for a large fragment of the language. I will present several example models and discuss interesting issues arising from the treatment of evidence in such languages.
Tuesday, Nov 30, 2004; 405 Soda Hall, 4:00-5:30pm
	Nov 30, 2004	4:00-5:30pm; 405 Soda Hall	Recent Research in Cross-language Document Search Fredric Gey University of California, Berkeley Abstract Cross-language document search research has been underway for more than 10 years now and while much progress has been made, certain research challenges remain. This talk will review recent research in Cross-language information retrieval, including the 2004 evaluation workshops: NTCIR for Asian language retrieval in Japan (http://research.nii.ac.jp/ntcir-ws4/index.html) and CLEF for European language retrieval (http://clef.iei.pi.cnr.it:2002/), as well as the U.S. DARPA “Hindi Surprise Language Exercise” of 2003. Topics to be covered include: Language-specific processing (stemming, segmentation, stop-words) Word decompounding for German Translation disambiguation for bilingual dictionaries Parallel corpora induced lexicons Web corpora usage for out-of-vocabulary translation Special retrieval tasks (Patent Retrieval, Cross-language question answering Geographic information retrieval Challenges of less-commonly taught languages The road ahead in cross-language information retrieval research Presenter: Dr. Fredric Gey has been doing research in cross-language information retrieval since 1998. He and his associates have participated in every cross-language information retrieval evaluation in the United States, Japan and Europe. Currently he is working on retrieval (including geographic information retrieval) of Russian language corpora and other digital objects. Dr. Gey co-chaired the English-Arabic retrieval evaluation track at the TREC conferences in 2001 and 2002. He co-chaired a workshop on “Cross-language Information Retrieval Research: The Road Ahead” at the ACM SIGIR-2002 conference in Finland. He is co-author of the entry on “Multilingual Information Retrieval” in the Encyclopedia of Library and Information Science and co-editor of a forthcoming special issue on Cross-Language Information Retrieval of the Information Processing and Management Journal.

Notify the calendar administrator of a change to an existing EE or CS calendar entry.

Powered by WebEvent (tm).