Alan Biermann

• Office: D213 LSRC
• Phone: 660-6539
• E-mail: awb@cs.duke.edu

• Office: D105 LSRC

Meeting Times

MWF 10:30-11:20am D243 LSRC

Course Goals

1. To obtain a grounding in the fundamentals of computational linguisitics by covering most of a standard text in the field, and

2. To have a research experience in computational linguistics by studying frontier papers in the field and facing an unsolved problem or an unknown area in the field and discovering its basic parameters and characteristics.

Method

The course will advance through most of the chapters of the main textbook (by James Allen) and then will cover a series of research papers selected by the instructors. The areas of concentration for the advanced papers will be speech interactive systems, dialogues systems, multimedia systems, and knowledge representation. Class presentations will be partly by lectures from the instructors and partly by student presentation. Quizzes will be given each Wednesday over the material covered up to that time in the course. (In some cases, homework problems may be given in place of quizzes.)

Students will each select a research topic early in the course and will spend the whole semester researching it. (Students may work in teams of two if they can divide the task in an acceptable way.) The instructors will suggest a number of projects that are both doable in the alloted time and that will have the potential of advancing the state of the art in the field. Each student will meet regularly with the instructor to discuss progress on the topic and to guarantee a successful conclusion at the semester end. The instructor will give one or more lectures near the end of the course on how to write a technical paper for publication and each student will write such a paper. Students will then be instructed on how to review a paper submitted for publication and each student will review papers written by other members of the class.

Grading

The grade will be based on performance in the above two areas (reading material and project) with roughly half of the grade being based on each area.

Textbook

James Allen, Natural Language Understanding, Benjamin/Cummings Publishing Co., 1995. (This should be available from the bookstore.)

Examples of some papers that will be covered:

• Barbara Grosz and Candice Sidner (1986), "Attentions, Intentions, and the Structure of Discourse," Computational Linguistics, 12(3), 175-204.

• James Allen et al. (1995), "Trains," Tech. Report.

• Raymond Perrault and James Allen (1980), "A Plan Based Analysis of Indirect Speech Acts," Computational Linguistics, 167-182.