GSLT: Natural Language Processing (Level 1)

The aim of this course is to give a research-oriented overview of natural language processing focusing on the following two questions:

What are the standard methods used in the field?
What are the current research problems?

The course is aimed both at students with limited knowledge of the field, for whom it is compulsory within GSLT, and at students with a more extensive background in natural language processing, who will be expected to take more active part in the discussion of current research. In this way, the course is meant to contribute to the common platform for students with different backgrounds within GSLT.

NB: The official language within GSLT is English but we can decide to have the lectures in Swedish instead, provided of course that all participants are comfortable with this. In any case, participants are free to formulate their contributions to discussions, whether oral or written, in any language that can be understood by the other participants (which in most circumstances means Swedish or English).

Part 1: Introductory Lectures (Göteborg)

	Date	Time	Room	Contents	Slides	Reading	Teacher
1	27/1	8-10	C430	Introduction and overview	Introduction	J&M 1	JN
2	27/1	10-12	C430	Words: morphology, n-grams	Words	J&M 2-3, 6	JN
3	28/1	8-10	C430	Syntax: tagging, parsing	Tagging Parsing	J&M 8-12	TL
4	9/3	8-10	C430	Semantics: semantic analysis, disambiguation	Semantics	J&M 14-17	JN
5	9/3	10-12	C430	Pragmatics: discourse, generation, translation	Pragmatics	J&M 18, 20-21	BG

Teachers: Lars Borin (LB), Barbara Gawronska (BG), Torbjörn Lager (TL), Joakim Nivre (JN)

Part 2: Distance Study Periods

The distance learning part of the course consists of four study periods of three weeks each:

Words [2 - 20 February]
Coordinator: Lars Borin
Syntax [23 February - 5 March, 15 - 19 March]
Coordinator: Torbjörn Lager
Semantics [22 March - 16 April]
Coordinator: Joakim Nivre
Pragmatics [19 April - 8 May]
Coordinator: Barbara Gawronska

During each of these study periods, students will be engaged in two kinds of activities:

A practical assignment, dealing with the implementation and/or evaluation of standard methods in the respective areas. (Students with different background knowledge may do different assignments here.)
A theoretical discussion of current research problems in the form of a chat session, based on the reading of recent articles and with an introduction by the respective teacher sent in advance by e-mail.

In addition to the practical exercises and chat discussions, students will be expected to produce a term paper where they discuss a research problem of their own choice in relation to other areas of natural language processing. (``How does my research problem fit into the world of natural language processing?'')

Part 3: Closing Seminar (Växjö)

The course will end with a seminar, where students will present their papers. The seminar will be held at Växjö University, 2-3 June 2004.

Course Requirements

In order to pass the course students must fulfil the following requirements:

Complete four practical assignments (one for each study period)
These assignments can be done individually or in groups of two (possibly three) students. Assignments should be completed by the end of each study period and the solutions made available on the group's web page. If an assignment is not completed on time, the group will be given an extra assignment which has to be completed before the end of the following study period. (The end of the following study period is also the deadline for assignments that have to be resubmitted.) Teachers are required to return assignments within one week.
Participate actively in the theoretical discussions (chat seminars)
Students are expected to contribute to discussions in each of the four study periods (and should notify the responsible teacher if they are unable to do so). Students who fail to meet this requirement may be given another assignment instead.
Write and present a term paper (5-10 pages)
The set topic for this paper is to discuss the place of your own research problem within the larger field of NLP and to relate it to other problems and methods. (Other topics may be chosen after approval from the main responsible teacher.) Papers have to be finished by 23 May and will be discussed at a seminar in Växjö 2-3 June, where each author has to make a short oral introduction. Each paper will be read by at least two other students.

Reading List

Introductory lectures:

J&M = Jurafsky, D. & Martin, J. H. (2000) Speech and Language Processing. Prentice-Hall.

Study period 1 (words):

Karttunen, L., Chanod, J-P., Grefenstette, G. and Schille, A. (1996) Regular expressions for language engineering. Natural Language Engineering 2(4): 305-328.
Minnen, G., Carroll, J. and Pearce, D. (2001) Applied morphological processing of English. Natural Language Engineering 7(3): 207-223.

Study period 2 (syntax):

Eugene Charniak (1997) Statistical Techniques for Natural Language Parsing. AI Magazine.
Brill, Eric (1995) Transformation-Based Error-Driven Learning and Natural Language Processing: A Case Study in Part of Speech Tagging. Computational Linguistics, December 1995.
Steven Abney (1997). Partial Parsing via Finite-State Cascades. J. of Natural Language Engineering, 2(4): 337-344.

Study period 3 (semantics):

Pinkal, M. (1999) On Semantic Underspecification. In Bunt, H. and Muskens, R. (eds.) Computing Meaning. Kluwer, 33-56.
Ide, N. and Véronis, J. (1998) Word Sense Disambiguation: The State of the Art. Computational Linguistics,24:1, 1-40.

Study period 4 (pragmatics):

Not available yet.