Mark A. Mandel

at the University of Pennsylvania

What I do

I am the research administrator of the ITR Grant in Biomedical Information Extraction. In non-technical terms, we are analyzing biomedical research texts to develop computer programs to extract information from them and put it into databases that researchers can search without having to find and read all the texts. That's not complete, but it covers a lot of our work.

Some numbers from our annotation production system:

• recent overview of number of files at each stage of processing
• recent character and token counts for POS-tagged files, both with and without the extraneous parts of files (anything beyond the abstract title and body)

For annotators

I am now hiring and training annotators.

• Job description
• Introductory overview of the project for annotators
  • PowerPoint presentation
  • web page with frames

Finding me

Institute for Research in Cognitive Science
Suite 400A, 3401 Walnut Street
Philadelphia, PA 19104-6228
215 898-0328
mamandel@ldc.upenn.edu

Personal

My personal home page is here.

Some project documents

• Upcoming conferences that might be relevant for the project, compiled by Seth Kulick.
   
• Summary of the ITR/E oncology work: Paper submitted to HLT/NAACL 2004 (Human Language Technology conference / North American chapter of the Association for Computational Linguistics annual meeting). Pete White writes: "This manuscript has been submitted to [this workshop]. It is our information extraction group's latest summary of our work and thus represents the best current overall perspective of the project."
   
• Pete White's presentation (2003-11-04) on some of our work at HGVS 2003 (PowerPoint).
   
• CorefFest (2003-05-15): Like ChemFest (below), but dedicated to the concept(s) of coreference.
  • the summary distributed at the end of the meeting
  • the detailed notes taken by Scott Winters.
  
   
• ChemFest (2003-04-03): a day-long meeting to hammer out an initial definition of chemical entities for the purpose of automatic and human tagging. We referred to 40 MedLine abstracts supplied by Andy Schein:
  • 10 on inhibition of cytochrome P450
  • 30 drawn at random
  and later to 10 on protein-DNA complexes that Alex Vasserman brought.
  • Here's what we had at the end of the day.
  
   
• Resources for Biomedical Terminology and Ontology: a work in progress.
   

2004-11-09