Information extraction from text, Week 5

The solutions should be ready for inspection by Thursday 14.3.2002 (midnight).

Remember that always, if you are in doubt what you should do, you can ask Reeta or send a message to our newsgroup!!

1. Study the paper Soderland, Lehnert: Learning Domain-Specific Discourse Rules for Information Extraction (AAAI, 1995) and find answers to the following questions:

   • In this domain, what is the input for the discourse-level processing like?
   • Find a few examples of discourse-level processing.
   • What kind of features are used to learn the processing rules? That is, what kind of information is given to the learning program (here ID3).

   A longer version of the ideas is available: Soderland, Lehnert: Wrap-Up: a Trainable Discourse Module for Information Extraction (Journal of Artificial Intelligence Research, 1994).

2. Study the paper Soderland: Learning to Extract Text-based Information from the World Wide Web (KDD, 1997), and try to apply its ideas for extracting humidity (kosteus) and wind speed (tuulen nopeus: "tuuli puhaltaa nopeudella x m/s") from the EKarjala-portaali . There are weather reports for Lappeenranta (lentokenttä, Kimpinen), Joutseno, and Parikkala available. If you have language problems, do not hesitate to ask more advice.