As some of you know, on Thursday June 2, 2005 I gave a talk and visited at Thomson West Publishing in Eagan, Minnesota (a suburb of the Twin Cities). This is one of the larger and historically more important companies in the Twin Cities, and is one of the largest employers of high tech employees in the Twin Cities.
Thomson is a large conglomerate of publishing organizations, and West Publishing was for many years a stand-alone company that publishes legal materials (books, online search services, etc. etc.) West was founded in 1872, and I think has always been in the Twin Cities (mostly in St. Paul). At some point in the not too distant past (perhaps in the 90s) Thomson acquired West, and became a much larger company as a result. I spoke with the research group there, which seems to have a general interest in doing information extraction from legal documents. An example publication which describes some of that work is : http://portal.acm.org/citation.cfm?id=964770 The law is actually a fairly big area of interest in AI and NLP in general, as witnessed by organizations such as this: http://www.iaail.org And here.s kind of an interesting article about the possible future of AI and the Law: http://www.economist.com/science/tq/displayStory.cfm?story_id=3714082 The legal profession has a tremendous amount of text that must be dealt with, including all existing laws, previous court decisions, dictionaries, guide to legal procedures, etc. Since the law changes constantly and since there is always new litigation, the amount of text grows quickly. This motivates much of the work at ThomsonWest, at least with respect to NLP and IR, I think. Among their goals are to develop tools that allow legal professionals to search through these huge amounts of data in a reasonably short amount of time, and to find what they are looking for conveniently. There is also the issue of managing such huge amounts of information, so that involves summarization and organization of text. So, I think it.s fair to say that ThomsonWest does or could potentially deal with many of the well known problems in NLP. They have a very famous tool in the legal community known as WestLaw, which is used by lawyers and judges around the country to access legal information. You can find info about WestLaw at: http://www.westlaw.com This is available via a web interface, so I imagine that many of the issues involved in Web search are present here, except in a more specialized domain. What follows are a few notes about the various people I met with. There is no claim that these summaries are complete by any means, but hopefully I have spelled the names correctly and made it possible for you to find some information about each person. The research group is led by Khalid Al-Kofahi, who has a PhD from RPI. A number of his recent publications are listed here, which include other Thomson folks as co-authors, and probably give a good general feeling for the nature of the work they do: http://www.informatik.uni-trier.de/~ley/db/indices/a-tree/a/Al=Kofahi:Khalid.html I got the impression that Khalid was very interested in classification, as exemplified by the following paper: http://portal.acm.org/citation.cfm?id=502603 Peter Jackson is the Chief Research Scientist and Vice President of Technology for Thomson Legal and Regularity. He has broad interests in NLP and AI, and has for example published a textbook on expert systems. (He is one step (at least) above Khalid in the Thomson hierarchy.) You can find his home page with information about his background and interests here: http://members.aol.com/JacksonPE/ Jack Conrad is a Senior Research Scientist and has been my main point of contact (he.s the person who arranged my visit). His home page is here, which includes a number of papers and other research related information: http://home.earthlink.net/~jackgconrad/ Jack is very interested in clustering, as described in a 2005 tech report: https://wwws.cs.umn.edu/tech_reports/index.cgi?selectedyear=&mode=printreport&report_id=05-016 Other researchers I spoke with at some length included Christopher Dozier, Isabelle Moulinier, and Tonya Custis, although I could not locate home pages for them. There were also a few other folks I visited with, although I.m ashamed to say I can not remember their full names well enough to include here. We talked a good deal about the problem of name disambiguation/discrimination. This is a big deal in the legal world of course, as you have many lawyers, judges, defendant, etc. many of whom will have the same names. Keeping track of which John Smith is the one accused of murder versus which one is a lawyer in Kansas City is obviously an important issue. In addition, another interesting problem is that of creating indexes of companies that are involved in various kinds of legal proceedings. This is meant to create a tool for law firms that wish to market their services. For example, if you are a lawyer that specializes in anti-trust law, it would be nice to know all the companies that are involved in anti-trust litigation now and in the recent past. There is also an interest in the problem of conceptual search, that is rather than having to look for exact keyword matches, be able to organize and find information based on concepts. So if you search for .civil rights violations. you might find documents about .racism. and .sexism. even if they didn.t contain the term .civil rights violation.. (That.s definitely a layman.s example, I am sure there are better legal examples!) In any case, I found this to be a very interesting place with a lot of interesting work going on. I am sure that this summary has missed a lot, but hopefully it gives a sense of the place and the people I met. Let me know if you have any further questions, and if you are looking for a job you should certainly consider Thomson West as a place to apply (assuming you have some interest in any of the above that I describe!) Enjoy, Ted -- Ted Pedersen http://www.d.umn.edu/~tpederse ------------------------ Yahoo! Groups Sponsor --------------------~--> Take a look at donorschoose.org, an excellent charitable web site for anyone who cares about public education! http://us.click.yahoo.com/1gzaND/8WnJAA/HwKMAA/x3XolB/TM --------------------------------------------------------------------~-> Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/nlpatumd/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
