Posts Tagged “text-similarity”

I had the chance to attend this interesting webinar hosted by Pubget, a new search engine for life-science PDFs. The webinar was held on Friday, December 11, 2009 (you can catch the recording here). There were 160 attendees and the GoToWebinar tool enabled live interaction with the speakers.

The webinar meant to have speakers who are experts in their areas and to cover different segments dealing with searching, analyzing, and reusing scientific articles. The webinar was moderated by Ryan Jones, President of Pubget, and the speakers represented:

Peter Binfield talked about his experience with PLoS One as a journal established in the digital era, and all of its content is digital. He was much concerned with how to monitor the “reuse” of an article and the tools incorporated in PLoS to achieve that. PLoS uses multi-dimensional, article-level metrics rather than a monolithic system like impact factor. PLoS metrics system enables every one to know the exact usage of an article, downloads and views. PLoS also enables commenting, rating, discussing, selecting a part/line and writing a note about it, sharing/bookmarking, and showing trackbacks to blogs and citations.

Marcus Banks said that the digital “libraries” are still in need of a librarian to analyze, organize and link publications. He also talked about the need of a tool that enables researchers to highlight only the parts of a publication that they need, instead of consuming time reading through the whole publication. He talked about sharing tools like: Zotero, Mendeley,, RefShare, CiteULike, and Pubget.

Representing the end-users was Ansuman Chattopadhyay on the stand. His presentation was entitled: “Beyond PubMed: Next generation literature searching”. With PubMed, it’s difficult to narrow down your search and reduce the number of the results/hits, but this could be achieved by the newer Google-like tools such as:

  • GoPubMed, which gives the users suggestions as they are typing
  • Novoseek, which categorizes search results into: diseases, pharmacological substances, genes/proteins, procedures, organisms, etc.

and text-similarity tools like:

  • eTBLAST, a web server to identify expert reviewers, appropriate journals and similar publications (the paper)
  • JANE, Journal/Author Name Estimator
  • DeepDyve

One point I didn’t get is the need of a “daily journal of negative results”.

Ramy Arnaout presented Pubget as a search tool that is:

  • like an on-the-web Acrobat Reader (the search results are the PDFs of the papers)
  • able to deliver science at speed
  • legal and free, as researchers use their institution’s license to get to all publications including the non-open-access ones
  • user-friendly, as a user chooses from a list of publications a paper that opens in the same window

The concerns that all four speakers expressed at the end of the webinar were mostly:

  • How to achieve the balance between delivering science and preserving copyrights, a problem that is being partly solved by Open-Access journals.
  • How to tell the end-user what is related to his/her field.
  • Although everything is “online”, the challenge is how to get to it and use it.
  • How to interact with the end-users and make them discover the tools/features of search engines, this can be solved by workshops and tutorials.

I do thank Pubget for giving me the chance to attend this very informative webinar by making it  freely available.

Edited on Dec 22, 2009 09:31 p.m. CLT

Tags: , , , , , , , , , , , , , , , , , , , , , , , , , , ,

Comments 3 Comments »