Wednesday, April 20, 2011

Paper Reading #24

Comments:
Comment #1
Comment #2

Reference:
Title: Finding Your Way in a Multi-dimensional Semantic Space with Luminoso
Author: Robert Speer, Catherine Havasi, Nichole Treadway, Henry Lieberman
Venue: IUI 2009/2010
Summary:
Luminoso is an interactive system that helps researchers form correlations between words. It reads in text documents, then the user "grabs" a point of reference. From that, it allows the researchers to rotate their viewpoint in N-dimensional space. Thus, Luminoso is a form of data mining. Other similar algorithms that are automated will create useless or irrelevant relations between words, but since a user guides Luminoso, these are eliminated, giving the system a sort of "common sense." The group believes Luminoso could be useful in the creation of semantic networks, such as biological or medical informational resources, "by providing a visualizer which shows the layout, focus, and coverage of the developing resource." The picture below is an example of a similar program.

Discussion:
Another short paper makes me very happy. I think my luck runs out after this one though. Nonetheless, this paper I thought was very confusing. I didn't feel that it was particularly well organized which made it difficult to follow. I also didn't think the concepts were well explained until the end of the paper, requiring me to try and read the paper a second time. The concept seems alright, but again poorly explained. I don't think this will ever amount to anything because of such a small target market.

2 comments:

  1. I have to agree with you that I was looking forward to a short paper, but this one was actually harder to read than the longer papers we have had.

    I do have to say that it is difficult to organize a bunch of word data, and while this is the wrong way to go about it in my mind, it is an interesting idea nonetheless.

    ReplyDelete
  2. I have had a few papers that were difficult to read and I found I didn't care for the ideas they presented. I wonder if bad papers generally have bad ideas or if the bad paper makes us think the ideas are bad.

    ReplyDelete