Generation of Referring Expressions in Natural Language

a course in the Sino-European Winter School in Logic, Language and Computation (SELLC), Guangzhou, Dec. 2010

by Kees van Deemter, University of Aberdeen

  1. Reading relevant parts of "Computational Generation of Referring Expressions: A Survey" would be useful course preparation. Here is a near-final version, revised from the one that was used at the summer school. The final version of this survey, by Emiel Krahmer and Kees van Deemter, is appearing in Computational Linguistics 38(1), March 2012, pp.173-218. It can be downloaded for free from the journal's web site.
  2. Here are drafts of lecture slides. We won't have time to cover everything during the course, but the first five items are definitely on the menu. A choice between the other items will be made on the fly.

    1. Introduction to the course: slides in ppt

    2. General introduction to Natural Language Generation: slides in ppt

    3. The Incremental Algorithm (1): the algorithm. slides in ppt

    4. The Incremental Algorithm (2): testing the algorithm. slides in ppt

    5. Referring to sets (1): general principles. slides in pdf

    6. Referring to sets (2): aiming for coherence. slides in ppt

    7. Referring to sets (3): avoiding surface ambiguities. slides in ppt

    8. Referring to sets (4): concluding remarks. slides in ppt

    9. Allowing hearers to find the referent (1): an algorithm for referring to items in hierarchical domains. slides in ppt

    10. Allowing hearers to find the referent (2): general remarks. slides in ppt

    11. Allowing hearers to find the referent (3): psycholinguistic studies. slides in ppt

    12. Generating vague descriptions (1): the algorithm. slides in pdf

    13. Generating vague descriptions (2): Why be vague? slides in ppt

    14. Reference and quantification: an account based on Description Logic. slides in ppt. And the same slides in pdf (in case of problems with the rendering of symbols in ppt)

    15. Open Questions slides in ppt