Dependency Parsing (Summer 2021)
Quick Links
General Information
Instructors: | Kilian Evang |
Jakub Waszczuk | |
Lectures: | Tuesday, 14:30 – 16:00, online |
Labs: | Friday, 10:30 – 12:00, online |
Rocket.Chat channel: | https://rocketchat.hhu.de/group/dependency-parsing-2021 |
Course web page: | https://kilian.evang.name/teaching/hhu-dp-su21/ |
(This web page, which will be updated throughout the course.) | |
Languages: | English and German |
Course Description
The task of syntactic parsing consists in finding the structure of a natural language sentence. Just as there are two main approaches of representing such structures – namely, constituency- and dependency-based structures – there are also two main approaches of syntactic parsing, adapted to handle constituency trees and dependency graphs, respectively. Since dependency parsing adopts a different view on how structure of languages should be represented, the methods used for dependency parsing are often different from those used for constituency parsing.
Dependency parsing can be performed either using linguistic descriptions (grammars) or using machine learning (data driven). This course presents the methods used in data-driven dependency parsing. These methods fall in two main families, the first one being based on transition systems, the second one on graph theory.
Requirements
- BN: homework exercises (> 50% correct for bachelor students, > 60% for master students); preferably in groups (up to 3 members)
- AP: homework exercises (see BN) + written exam (with extended exercises for master students)
Schedule
Week 1 | Introduction and overview |
Exercises (latex template, annotations) | |
Week 2 | Basics of dependency parsing (constituency to dependency conversion example; examples of dependency graph properties) |
Exercises (latex template, solutions) | |
Week 3 | Grammar-based approaches to dependency parsing (lecture notes) |
Exercises (latex template, solutions) | |
Week 4 | Introduction to data-driven dependency parsing |
Exercises (.xlsx template, possible solution) | |
Week 5 |
Transition-based dependency parsing (part 1) Projective shift-reduce (a.k.a. arc-standard) parsing; see also Nivre (2008) for correctness proofs and complexity analysis
|
Exercises (latex template, solutions) | |
Week 6 | Arc-eager parsing, non-projective parsing |
Exercises (latex template, solutions) | |
Week 7 |
Transition-based dependency parsing (part 3) Dynamic oracles
|
Exercises (latex template, solutions)
|
|
Week 8 | Implementation of a transition-based parser (part 1) |
Exercises (blog post, source code, data, solution) | |
Week 9 | Implementation of a transition-based parser (part 2) |
Exercises (solution) | |
Week 10 | Implementation of a transition-based parser (part 3) |
Exercises (starter code, solution) | |
Week 11 | Chu-Liu-Edmonds’ algorithm |
Exercises (solution) | |
Week 12 | Eisner algorithm |
Exercises (latex source, solution) | |
Week 13 |
Graph-based dependency parsing (part 3) |
Exercises (solution) | |
Week 14 | Revision for the exam |
Exercises from previous years | |
Week 15 | Mock exam |
Solution | |
2021-07-30 10:30–12:00 |
Online exam Leitfaden zur Durchführung von Online-Klausuren The exam is an open book exam, i.e., all course materials, published papers, textbooks, etc., may be used. It is not permitted for participants to communicate with each other during the exam. It is not permitted to use interactive resources such as online parsing GUIs. The exam questions will be posed in English. You can answer them in English or German. If you need a work station at uni for the online exam, please email me ASAP! |
Acknowledgments
Thanks to Simon Petitjean, Andreas van Cranenburgh, and Rafael Ehren for providing their teaching materials, as well as to Sandra Kübler, Ryan McDonald, and Joakim Nivre, authors of the various tutorials (ACL 2006, ESSLLI 2007, EACL 2014) on which this course is based.