Whither DOP?

 

ILLC Workshop, Friday 9 December 2005

Plantage Muidergracht 24

University of Amsterdam

Room 3.27

(NB: if room 3.27 turns out to be too small, we will move to T.201, Sarphatistraat 104, 
two minutes walking from the ILLC main venue, see map)
 

 

Organizers:       Rens Bod, Remko Scha and Khalil Sima`an

Sponsors:         NWO, ILLC

 

Now that the DOP approach has entered its 15th year from its original conception, the time has come to critically look back and think afresh about the future. What went astray with DOP's statistical estimator and has it been solved? What are the problems related to DOP's productive units, and how can they be redressed? What else is still to be done in DOP? And how far can the model be stretched?

 

DOP has led to statistical extensions of Lexical-Functional Grammar, Head-driven Phrase-Structure Grammar and Tree-Adjoining Grammar. It has been applied to machine translation, speech understanding, music analysis, problem-solving and equational reasoning. Can it be extended to other fields of cognition? What are the counter-examples that DOP has to face? And what is its relation with frameworks like Case-Based Reasoning and Probabilistic Inductive Logic Programming?

 

This workshop aims to provide an informal discussion platform on DOP and is open to everyone. Each talk will discuss one or more open problems and will be followed by ample discussion. The workshop is also meant to give an overview of the projects and/or recently founded research groups on DOP at the University of Amsterdam, the University of St Andrews, at Dublin City University and the University of Essex.

 

For more information on DOP go to the DOP homepages.

 

For the abstracts received so far, click here.

 

For a global map of the venue including hotel and restaurant, click here.  For a more detailed map of the workshop venue (look for Plantage Muidergracht 24), click here.

 

If you want to meet (some of) the speakers before the workshop, come to Café De Jaren (see map), on Thursday 8 December, 9.30 pm.

 
                                              Time Schedule
 
9.45-10.00: Coffee
 
10.00-10.20: Remko Scha, “DOP as a Cognitive Model
10.20-10.40: Discussion
 
10.40-11.00: Andy Way, “DOP Models of Translation: Theory & Practice
11.00-11.20: Discussion
 
11.20-11.35: Coffee
 
11.35-11.55: Khalil Sima'an, “Estimators for Generative All-Subtree Models
11.55-12.15: Discussion
 
12.15-12.35: Evita Linardaki, “Reconsidering DOP Probabilities
12.35-12.55: Discussion
 
13.00-14.15: Lunch
 
14.15-14.35: Jelle Zuidema, “Data Oriented Language Learning: Weight 
Estimation and Structure Search in the Unsupervised Learning of STSGs
14.35-14.55: Discussion
 
14.55-15.15: Mary Hearne, “From tree-based parsing and translation to 
LFG-DOP and LFG-DOT: what are the challenges?
15.15-15.35: Discussion
 
15.35-15.50: Coffee/Tea
 
15.50-16.10: Detlef Prescher, “Treebank Grammars and Other Infinite 
                                   Parameter Models”
16.10-16.30: Discussion
 
16.30-16.50: Rens Bod, “Whither DOP? Unifying with other modalities or 
integrating with unsupervised learning?”
16.50-17.10: Discussion
 
17.10-17.20: Tea with snack
 
17.20-18.00: General discussion
 
19.00: dinner at SSK Restaurant (for invitees only)
 
Since space is restricted, please send an email to rens@science.uva.nl if 
you want to join the Whither DOP-workshop.