Computational Pragmatics
Lecturer: Raquel
Fernández (Institute for Logic, Language and Computation)
Teaching assistant: Julian Schlöder
Timetable: Tuesdays and Fridays 13-15h; room SP G3.05 (see https://rooster.uva.nl)
Overview: This year the course will be dedicated to dialogue modelling
(to models of language as it is used in actual conversation). This website will be updated
throughout the course with the contents of each session. This information may be tentative and it is subject to change. Please check this website regularly.
Prerequisites:
There are no formal prerequisites, but I'll assume some basic
knowledge of semantics/pragmatics of natural language. Basic programming
skills are useful.
Evaluation:
The first part of the course will consist of
lectures and discussions of research papers; everybody is expected
to play an active role in class. In the second part of the course you should work on a team project of your choice (see
below). You will be asked to write
a short paper by the end of the course. This paper will account for most of the
final grade (70%). The rest will correspond to exercises and class discussions.
01 Sep
Brief overview of the course and introduction to turn taking. [slides]
Some references with overviews of the field (I recommend you to read
at least one of them):
To do:
- Read at least one of the two papers we will discuss on Friday 4 Sept. You don't need to undesrtand every technique being used in the paper, but you should be able to asnwer, in a concrete way, questions such as: what is the main aim
of the work reported in the paper? which methodology has
been used? What are the results? Also: what did you find exciting about the paper? what was not convincing?
- Homework #1: due on Tuesday 8 Sept, 13h (bring a printout to class). [transcription guidelines]
04 Sep
Discussion of the following two research papers on turn taking:
08 Sep
Speech act theory and dialogue acts. In the last part of the lecture we discussed inter-annotator agreement. [slides]
To do:
- Read the paper we will discuss on Friday 11 Sep on recognizing dialogue acts in Wikipedia. Be ready to discuss it.
- Homework #2: due on Friday 18 Sept, 13h (bring a printout to class and send it by email).
11 Sep
Discussion of the following research paper on dialogue act recognition:
Introduction to grounding. [slides]
To do: Read the paper we will discuss on Tuesday 15 Sep related to decision-theoretic models of grounding. Be ready to discuss it.
15 Sep
More on grounding: grounding in referential tasks; decision-theoretic models. [slides]
Recommended reading: Clark & Brennan (1991) Grounding in Communication.
Discussion of the following paper:
18 Sep
Homework #2 is due at 13h. Bring a print out to class and send it also by email to Julian (including code).
Raquel is away. This lecture will be given by Julian. The lecture will include discussion of the transcription exercise. [slides]
22 Sep
Coordination and alignment. [slides]
To do:
- Read fully one of the papers we will discuss on Friday 25 Sept, but check the alignment measures used in both papers.
- Get ready to start discussing your project ideas on Friday.
- Homework #3: due on Friday 2 October, 13h. [optional supplementary material: starting script, coordination markers]
25 Sep
Discussion of the following papers related to alignment:
29 Oct
Brief introduction to dynamic semantics for dialogue.[slides]
Background reading: first
chapter of The Interactive Stance: Meaning for Conversation by Jonathan Ginzburg.
02 Oct
This class is cancelled. I recommend you to use the time to work on your project outline.
Homework #2 is due at 13h -- send it by email to Julian.
Project proposals due (by midnight). Check the guidelines and send an email to Raquel with your project outline in the body of your email.
Project discussion meetings
Meetings to discuss ongoing project work (Raquel's office at the ILLC, F1.07):
- First round:
Thursday 8 Oct: 15h Roger, Laurato & David
Friday 9 Oct: 11h Sharon & Elise, 12h Nina & Thom, 14h Omer.
- Second round: bring a draft of your paper (minimum two pages long)
Friday 16 Oct: 11h David, Roger & Laurato, 12h Nina & Thom, 14h Sharon & Elise, 15h Omer.
Project presentations
Project pesentations: Wednesday 21 Oct, 10:30-13h, ILLC seminar room F1.15 (opposite Raquel's office). Bring your own laptop please.
- Schedule: The time slots are adapted to the number of team members per project. Presentation time should be shared as evently as possible amongst team members. Make sure you leave between 5 and 10 minutes for questions and discussion.
- 10:50 Introduction: please arrive on time.
- 11:00 Roger, Laurato, and David on dialogue act tagging with utterance embeddings.
- 11:30 Omer on discourse markers in Dutch and English.
- 11:50 Elise and Sharon on linguistic coordination and disagreement.
- 12:15 break
- 12:25 Nina and Thom on linguistic coordination and comment rating.
- 12:50 Remarks on guidelines for final paper.
- Evaluation: We will grade your presentation according to
these criteria. The presentation is
worth 10% of your final grade. Your project will most likely not be finished by the time of the presentation. That's of course fine. Indicate clearly what is yet to be done and take the chance to get feedback from your fellow students.
Final papers
- Submission deadline: Monday 26 October, midnight (submit on time! 1.5 points less for every day of delay). Email your final paper to Raquel. If you submit code or datasets (recommended but not part of your grade), provide suitable URLs (do not attach them to your email please).
- Format: Your paper must be formatted using the style files from the latest ACL conference (you can find them here, scroll down to FORMAT and ACL-ICNLP 2015 Style Files).
- Length: between 6 and 8 pages, including references (these length restrictions are strict: 0.5 points less for not adhering to them.)
- Evaluation: I will evaluate your final paper following criteria similar to those typically used to evaluate papers submitted to an international computational linguistics conference (obviously adapted to the circumstances of a Master's course). The evaluation criteria are described here. Make sure you read these criteria when you start to plan your paper and re-read them again before submitting.
Guidelines for projects
[I will keep updating this section with information about the projects]
Projects should be done, ideally, in pairs. In very speacial circumstances, I may consider individual projects or larger teams. Your project can be about any topic connected with the contents of the course (i.e., generally about dialogue modelling). For inspiration, here are a couple of project papers from previous years (but bare in mind that these were individual projects, so the team projects this year should be a bit more substantial):
You need to talk to me about your project ideas before Friday 2nd of October. By that day, your project proposal needs to be submitted for final approval. This proposal must include:
- Team members.
- The goal of your project (what problem / phenomenon are you going to address? try to be as explicit as possible here).
- The kind of methodology you plan to employ and the data you will use (note that you may have to adapt your goal to the kind of data available).
- A sketch of the work plan (how will you get started? what steps do you forsee? ask yourself whether your plan is feasible given the timeframe).