|
PROXIMITY 4.3 released (November 15, 2007) The latest, open-source version of KDL's Proximity software for relational knowledge discovery is available for download.
Matthew Rattigan gives talk at Microsoft Research (June 26, 2007): Matthew Rattigan gave an invited talk at Microsoft Research in Redmond, WA. The talk focused on using Network Structure Indices [PDF] to efficiently search and cluster graphs [PDF].
Matthew Rattigan accepts summer internship at Comcast Interactive Media Labs (June 13, 2007) He'll be working on applications of relational learning to an entertainment web portal.
KDL predicts NFL playoff teams (September 22, 2006) Andrew Fast and David Jensen developed a statistical model to predict which teams will make this year's National Football League playoffs. The model is based on analysis of the social network among NFL coaches. The predictions are available on Andrew's homepage. This work will be presented at this fall's AAAI Symposium on Capturing and Using Patterns for Evidence Detection. AAAI awarded Andrew a grant for travel expenses.
Jennifer Neville completes dissertation (August 01, 2006) Jennifer Neville successfully defended her dissertation titled Statistical Models and Analysis Techniques for Learning in Relational Data. The computer science department nominated her dissertation for the ACM Doctoral Dissertation Award. This fall, she joins the computer science faculty at Purdue University.
KDL students publish paper at KDD 2006 (July 17, 2006) At this year's International Conference on Knowledge Discovery and Data Mining, Matthew Rattigan, Marc Maier, and David Jensen will present their work on network structure indexes [PDF].
Andrew Fast attends Doctoral Consortium (July 17, 2006) Andrew Fast presented his work at the Eleventh AAAI/SIGART Doctoral Consortium in Boston, Massachusetts during the Twenty-first National Conference on Artificial Intelligence. His presentation is titled "Learning Models of Macrobehaviors in Complex Adaptive Systems." The AAAI and ACM/SIGART Doctoral Consortium provides an opportunity for a group of Ph.D. students to discuss and explore their research interests and career objectives with a panel of established researchers in artificial intelligence.
KDL at NESCAI 2006 (April 28, 2006) KDL member Lisa Friedland presented her work on "Predicting Protein Complexes Using Relational Features from Experimental Data" at the first gathering of the North-East Student Colloquium on Artificial Intelligence (NESCAI).
PROXIMITY 4.2 released (April 24, 2006)
The latest, open-source version of KDL's Proximity software for relational
knowledge discovery is available for
download.
KDL releases relational datasets (March 24, 2006
and later)
KDL has released several datasets for use in relational
knowledge discovery research. The datasets are XML files
formatted for direct import into Proximity 3.0 or later. See
the Data page for more
information.
Ross Fairgrieve joins Google (January 9, 2006) Ross Fairgrieve joined Google's New York office as a Software Engineer. Ross is a former member of KDL and recently received a Masters in Computer Science from UMass.
Brian Gallagher joins Lawrence Livermore (January 1, 2006) Former KDL member and recent UMass Masters degree graduate, Brian Gallagher has started work as a Computer Scientist in the Center for Applied Scientific Computing at Lawrence Livermore National Laboratory. He is working on the Complex Networks Project, developing algorithms and analysis techniques for complex networks (e.g, social and biological networks and the Web). Currently he is working on pattern matching in large graph datasets and hopes to apply this work to (1) improving pattern query performance in graph databases and (2) object consolidation.
PROXIMITY 4.1 released (December 15, 2005) The latest, open-source version of KDL's Proximity software for relational knowledge discovery is available for download.
Marc Maier joins lab (December 14, 2005) Marc Maier joined KDL as a Research Assistant. Marc recently graduated from the University of Massachusetts Amherst with a B.S. in Computer Science and Mathematics. Marc comes into the Department as a Bay State Scholar. As an undergraduate, he worked in the Mathematics Department on a mathematical biology project on tumor-induced angiogenesis.
KDL, NASD collaboration appears in the news (October 10, 2005) An article on KDL's collaboration with the National Association of Securities Dealers (NASD) appeared in the UMass Daily Collegian. The work was also mentioned in online news digests such as ACM Tech News and KDnuggets. Lab members Jennifer Neville, Özgür Şimşek and David Jensen applied the lab's knowledge discovery system to the task of identifying misconduct among securities brokers. When results were compared to a watch list procured by NASD analysts, the lab's system identified many of the brokers on the list as well as new ones. The results impressed NASD officials. John Komoroske, vice president of the NASD, says "That it performs as well as live examiners is fascinating."
Lab's research on 'six degrees of separation' phenomenon gains media attention (September 12, 2005) The Lab's research on finding short paths between nodes in a network, a problem at the heart of the 'six degrees of separation' social phenomenon, was covered in an article in the New Scientist, a popular science magazine, and Science Now, Science magazine's online news site. The research was also mentioned on the ACM's Tech News bulletin, AAAI's AI Alert, and The Engineer Online. Lab members Özgür Şimşek and David Jensen developed a novel algorithm for finding a node in a network using only local topological information. The research discussed in the article was published earlier this year at the 19th International Joint Conference on Artificial Intelligence [PDF].
Brian Taylor joins the lab (September 9, 2005) Brian Taylor joined KDL as a Research Assistant. Prior to joining KDL, Brian worked as a researcher for Institute for Scientific Research. Brian earned his B.S. in Electrical and Computer Engineering and M.S. in Electrical Engineering from the West Virginia University.
NASD publicizes collaboration with KDL (August 23, 2005) KDL's collaborative project with the National Association of Securities Dealers (NASD) was mentioned in their 2004 Annual Report. Page 13 of the report states: "In 2004, we continued to refine our use of data-mining techniques to prevent wrongdoing. We are also working on advanced pattern detection technology designed to give us an early warning or alert us to problems in firms on a near real-time basis. For example, in one initiative, in conjunction with the University of Massachusetts, we are analyzing data already available to us to create a statistical model that predicts which brokers warrant additional supervisory or regulatory scrutiny."
KDL students invited to Doctoral Consortium (August 23, 2005) Jennifer Neville and Özgür Şimşek presented their work at the Tenth AAAI/SIGART Doctoral Consortium in Pittsburgh, Pennsylvania during the Twentieth National Conference on Artificial Intelligence. Their presentations were "Structure Learning for Statistical Relational Models" and "Towards Competence in Autonomous Agents". The AAAI and ACM/SIGART Doctoral Consortium provides an opportunity for a group of Ph.D. students to discuss and explore their research interests and career objectives with a panel of established researchers in artificial intelligence.
KDL students publish papers at 2005 conferences (June 14, 2005) At this year's International Conference on Knowledge Discovery and Data Mining, Andrew Fast, Brian Levine, and David Jensen will present a poster on their work on creating social networks to improve peer-to-peer networking [PDF], and Jennifer Neville, Özgür Şimşek, David Jensen and three external coauthors will present an application of relational knowledge discovery to the problem of preventing securities fraud [PDF]. In a paper that will be published at the 19th International Joint Conference on Artificial Intelligence, Özgür Şimşek and David Jensen present a novel algorithm for finding a node in a network using only local topological information [PDF].
KDL collaborates with NASD to prevent securities fraud (June 14, 2005) Jennifer Neville, Özgür Şimşek and David Jensen applied relational knowledge discovery to help the National Association of Securities Dealers (NASD) prevent and discover misconduct among securities brokers. Their findings will be published at the 11th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining [PDF].
David Jensen collaborates with the Laboratory for Perceptual Robotics (June 14, 2005) David Jensen worked with Rod Grupen and Steven Hart on applying statistical relational learning to train a humanoid robot to perform tasks such as picking up heavy objects. Their work will be published in the Proceedings of the 2005 American Association for Artificial Intelligence (AAAI) Conference [PDF].
PROXIMITY 4.0 released (April 1, 2005) The latest, open-source version of KDL's Proximity software for relational knowledge discovery is now available.
PROXIMITY 3.1 released (September 15, 2004) The latest, open-source version of KDL's Proximity software for relational knowledge discovery is now available.
Amy McGovern joins faculty at University of Oklahoma (Sept 1, 2004) Former KDL postdoc Amy McGovern is now an Assistant Professor at the University of Oklahoma School of Computer Science.
PROXIMITY 3.0 released (April 15, 2004) The new, open-source version of KDL's Proximity software for relational knowledge discovery is now available.
Dan Corkill joins KDL (January 5, 2004) Dan Corkill joined KDL as a Senior Research Scientist.
Cindy Loiselle joins KDL (December 15, 2003) Cindy Loiselle joined KDL as a Technical Writer. Prior to joining KDL, Cindy served as a technical writer for several regional technology companies. She brings ten years' experience in software and hardware documentation to her work at KDL. Cindy earned her M.S. in Computer Science from the University of Massachusetts.
UMass awarded ITR grant from NSF (September 26, 2003) A proposal written by Andrew McCallum and David Jensen ("Unified Graphical Models of Information Extraction and Data Mining with Application to Social Network Analysis") was funded by the National Science Foundation through their Information Technology Research (ITR) program. The grant will provide nearly $3 million over 5 years to dramatically improve the state-of-the-art in our ability to data mine information previously locked in unstructured natural language text.
KDL announces open source version of Proximity (September 15, 2003) KDL announced that version 3.0 of its Proximity software will be released in April 2004 under an open-source license. Proximity is an integrated system for knowledge discovery in relational data sets, and it implements a variety of techniques introduced by members of KDL in the past two years.
Pippin Wolfe joins KDL (September 2, 2003) CS graduate student Alicia Peregrin "Pippin" Wolfe joined the laboratory as a graduate research assistant. Prior to joining KDL, Pippin was a research assistant in the Autonomous Learning Laboratory (ALL), to which she retains close research ties.
KDL members present seven papers at ICML and SIGKDD 2003 (August 24, 2003) Students and staff of KDL presented seven papers at two collocated conferences: the International Conference on Machine Learning (ICML-2003) and the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (SIGKDD-2003), and their associated workshops. The papers, all available from KDL's publications page, covered a wide variety of recent research in the laboratory, including new algorithms for relational knowledge discovery, statistical properties of relational data, technical critiques of potential applications, and our first-place KDD Cup entry.
KDL wins major data mining competition (August 24, 2003) KDL members won first place in the 2003 KDD Cup competition, a data mining competition held in conjunction with the Ninth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD-2003). KDL's winning team consisted of Amy McGovern, Lisa Friedland, Michael Hay, Brian Gallagher, Andrew Fast, Jennifer Neville and David Jensen. The UMass team competed in the open task (one of four tracks in the 2003 KDD Cup competition) which allowed contestants to define their own analysis tasks. Winners were selected by a panel of judges. The title of the UMass entry was Exploiting Relational Structure to Understand Publication Patterns in High Energy Physics. The paper is available for download.
Amy McGovern organizes space applications workshop (August 21, 2003) KDL Postdoc Amy McGovern co-chaired the ICML2003 Workshop on Machine Learning Technologies for Autonomous Space Applications.
David Jensen organizes relational learning workshop (August 11, 2003) KDL Director David Jensen co-chaired the IJCAI 2003 Workshop on Learning Statistical Models from Relational Data.
Sandi Harris Graves joins KDL (February 4, 2003) Sandi Harris Graves joined KDL as a secretary.
Özgür Şimşek joins KDL (December 18, 2002) CS graduate student Özgür Şimşek joined KDL as a graduate research assistant. Prior to joining KDL, Özgür was a research assistant in the Autonomous Learning Laboratory (ALL), to which she retains close research ties.
Patricia Riddle and Michael Barley visit KDL (September 21, 2002) Pat Riddle and Mike Barley from Auckland University's Department of Computer Science began a three-month term as visiting faculty in KDL.
Lisa Friedland, Brian Gallagher, and Michael Hay join KDL (September 2, 2002) Three new CS graduate students Lisa Friedland, Brian Gallagher, and Michael Hay joined KDL as graduate research assistants.
Amy McGovern joins KDL (June 2002) Amy McGovern, who had recently completed her dissertation in the Computer Science Department under the direction of Professor Andrew Barto, joined KDL as a Senior Postdoctoral Research Associate.
Agustin Schapira joins KDL (June 2002) Agustin Schapira joined KDL as a Senior Staff Programmer.
Knowledge Discovery Laboratory created (September 1, 2001) The Knowledge Discovery Laboratory began its official existence as a research group within the University of Massachusetts Amherst Department of Computer Science.
|