Foundations of Artificial Intelligence (FAI) Group

Seminar: Search Problems in Natural Language Processing

Basics. Seminar, 7 graded ECTS points.

The seminar will be run in a block format. There will be an initial meeting on Wednesday, October 26, 17:00--18:00.. All student presentations will be given on a single day after the end of term. A detailed schedule is given below.

All meetings will take place in room 3.06, Building E1 1. The seminar language is English throughout.

Your task will be to read and understand a piece of research, to write a summary paper in your own words, to give a presentation, and to provide detailed feedback for the paper and presentation of a fellow student.

All email interaction must be pre-fixed with "[NLP16]" in the email subject.

No plagiarism. It is Ok (and encouraged!) to use web resources to further your understanding of your assigned topic. However, it is inadmissible to use pieces of such material for your summary paper or presentation. Any plagiarism will result in disqualification from the seminar.

Content. Search problems, involving the exploration of a large combinatorial space of possible options, occur in several sub-areas of Natural Language Processing (NLP), including sentence generation, parsing, text-to-speech synthesis, and the quantification of linguistic distance. These search problems relate intimately to search problems considered in Artificial Intelligence and, in particular, in Automatic Planning. The seminar considers sentence generation, as search in a space of partial sentences, and parsing, as search in a space of matching grammatical expressions, where that connection is particularly direct as the search space is generated by a set of formal rules -- the grammar -- akin to the action descriptions generating transitions in planning. Connections between NLP and Automatic Planning have been made (and we consider some of these in the seminar), but the fields have largely developed independently and the working hypothesis of the FAI group is that many more opportunities for technology transfer exist.

Prerequisites. Participants must have successfully completed an introductory course in Artificial Intelligence. They should be familiar with automatic planning at least to the extent of the material covered in the Artificial Intelligence course. Successful participation in our Automatic Planning course will be an advantage, but is not absolutely necessary to follow the seminar.

Registration. The seminar has participation slots for 10 students. Registration for the seminar will be open from October 1 until October 24 (midnight). Please do not try to register ahead of time; we will only consider applications reaching us within the given time window!

To apply for registration, send an email to Álvaro Torralba. In the email, give a brief description of your relevant background. In particular, say whether you got a BSc and from which university, and describe previous lectures/seminars you completed in the areas of Artificial Intelligence, Automatic Planning, and Natural Language Processing. For each relevant course, state the grade you obtained in that course. Say a few words regarding why you are interested in participating in the seminar.

You will be notified by email on October 25, informing you whether or not you are registered.

Attention! As there always are more students than we have places, to be fair towards your fellow students please do only sign up if you're really interested in following the seminar through to the end. In particular, according to the new study regulations, you are only allowed to withdraw within three weeks after the briefing of the seminar, i.e., until November 16. Later withdrawal counts as "failed".

Grading. The final grading will be based on:

• The quality of the feedback you provide to other students.
• The quality of your summary write-up.
• The ability to stick to deadlines.
• The quality of your final presentation.
• The interaction during the final presenations.

Summary Paper. For the summary paper, you must use this tex template. Note in particular that you are required to read at least 2 related papers, for the related work section.The seminar paper should be about 4 pages long (not counting the literature list, and in the double-column format of the template). This is a rough guideline, not a strict rule. If you need, say, 5-6 pages to do your paper justice then definitely do so.

Schedule and Deadlines.

NOTE: If you have a problem with the date of the block seminar (presentation day), e.g. a close-by exam, please let us know AS SOON AS POSSIBLE. We will try to re-schedule.

• October 1--24: Apply for registration.
• October 25: Registration notification.

• October 26, 17:00--18:00: Initial meeting.

• October 31: Send a ranked top-5 list of the topics you would like to take, by email Álvaro Torralba. That is, send something like "7,3,8,5,1".

• November 2: Receive your assignment: A topic; a mentee student (to whom you will provide feedback, see the following deadlines); and a mentor student (who will provide feedback to you, see the following deadlines). The mentee/mentor assignment will be a "cycle" through each of the 3 topic areas as listed below: within each area, the student with topic 1 mentors the student with topic 2, who mentors the student with topic 3, who mentors the student with topic 1.

  Read the material associated with your topic carefully, and prepare an initial version of your summary paper, using the tex template given above.

• November 23-- December 7 (ca.): Make an appointment with your supervisor to discuss your paper: Give a brief oral summary of the paper (15--20 minutes); ask questions; do not bring the summary paper (yet).

  NOTE: This, and the following deadlines marked with "(ca.)", are meant as a guideline. You are required to do these things, but if you do them 3-4 days earlier or later, that is no problem.

• December 21 (ca.): Send your summary paper to your mentor student (cc supervisor).
• January 11 (ca.): Send feedback regarding the summary paper to your mentee student (cc supervisor).
• January 25 (ca.): Send revised summary paper to your mentor student (cc supervisor).

• January 27 (ca.): Send feedback regarding the revised summary paper to your mentee student (cc supervisor).
• February 1 (ca.): Send presentation slides to mentor student (cc supervisor).
• February 3 (ca.): Send feedback regarding the presentation slides to your mentee student (cc supervisor).
• February 10 (ca.): Send revised presentation slides to mentor student (cc supervisor).
• February 15 (ca.): Send feedback regarding the revised presentation slides to your mentee student (cc supervisor).

• February 28 (strict): Send final summary paper and presentation slides by email to Álvaro Torralba.
• March 3: Give a presentation (20 minutes talk, plus 10 minutes discussion) in the block seminar. Attendance to all talks is required. Please try to stick to the 20 minutes time slot for your talk; it should not be a lot shorter, nor a lot longer.

Topics. Each participant will be assigned one topic. The overall amount and difficulty of the material associated with each topic is roughly balanced.

Area 1: TAG Sentence Generation.

1. Sentence generation as a planning problem. Paper: Area 1 Topic 1. Mentors topic 2, is mentored by topic 3.
2. Results of planning for sentence generation. Papers: Area 1 Topic 2 a Area 1 Topic 2 b. Mentors topic 3, is mentored by topic 1.
3. Polarity filtering. Paper: Area 1 Topic 3. Mentors topic 1, is mentored by topic 2.

Area 2: CCG Chart Realization.

1. CCG Chart Realization. Paper: Area 2 Topic 1. For reference: Area 2 Topics 1-2. Mentors topic 2, is mentored by topic 3.
2. Efficient Realization in CCG. Paper: Area 2 Topic 2. For reference: Area 2 Topics 1-2. Mentors topic 3, is mentored by topic 1.
3. Dead-end pruning for realization in CCG. Paper: Area 2 Topic 3. Mentors topic 1, is mentored by topic 2.

Area 3: Search in CCG Parsing.

1. A* CCG Parsing. Paper: Area 3 Topic 1. Mentors topic 2, is mentored by topic 3.
2. A* CCG Parsing with Supertagging. Paper: Area 3 Topic 2. Mentors topic 3, is mentored by topic 1.
3. LSTM CCG Parsing. Paper: Area 3 Topic 3. Mentors topic 1, is mentored by topic 2.