rethink the structure and extend Part VII (Communication)
We have the time in the current schedule. This year, the course considerably thinned out in the last two weeks.
- There are currently two chapters and their content is a bit haphazard.
- The first (29) should be on language technology and language assistance systems (i.e. ngrams and co)
- the second (30) should be about NL Semantics and symbolic methods (all of the current 29.2/3 should be moved here).
- Chapter 29 (new) should be extended (a positive side effect is that I will learn more about this)
- Evaluation methods for IR (to drill in on information needs), talk about CLEF, NTCIR, ...
- Semantic Methods in IR (stemming, stop words, synonyms, ...)
- Text Classification (Ham/Spam)
- Information Extraction
- Machine Translation
- Chapter 30 (new) should also be extended.
- I would love to have a full pipeline from syntax (the current chapter on PCFG is good) via semantics construction (take a small fragment from LBS, where we do not need lambdas) and then do some disambiguation via reasoning. Then we can use this as a better intro/advertisemeht to LBS.
- Extend Speech Acts and do something with them (or leave them out completely)
- Extending the Grammar with Agreement?
- show how to implement PCFG in ProLog (at least for CFG it is very easy; they can just be compiled into clauses).
- maybe also do the semantics construction in Prolog
Edited by Michael Kohlhase