Synopsis

Advanced Semantic Representation ("SemRep")
Lectures: MW 3:30-4:45, ICC 108

Natural language is an imperfect vehicle for meaning. On the one hand, some expressions can be interpreted in multiple ways; on the other hand, there are often many superficially divergent ways to express very similar meanings. Semantic representations attempt to disentangle these two effects by exposing similarities and differences in how a word or sentence is interpreted. Such representations, and algorithms for working with them, constitute a major research area in natural language processing.

This course will examine semantic representations for natural language from a computational/NLP perspective. Through readings, presentations, discussions, and hands-on exercises, we will put a semantic representation under the microscope to assess its strengths and weaknesses. For each representation we will confront questions such as: What aspects of meaning are and are not captured? How well does the representation scale to the large vocabulary of a language? What assumptions does it make about grammar? How language-specific is it? In what ways does it facilitate manual annotation and automatic analysis? What datasets and algorithms have been developed for the representation? What has it been used for? In Spring 2017 the focus will be on the Abstract Meaning Representation (AMR); its relationship to other representations in the literature will also be considered. Term projects will consist of (i) innovating on the representation's design, datasets, or analysis algorithms, or (ii) applying it to questions in linguistics or downstream NLP tasks.

Credits: 3

Prerequisites: COSC/LING-572 (Empirical Methods in Natural Language Processing) or LING-367 (Computational Corpus Linguistics)

(For administrative reasons, registration is divided into two sections: LING-672 and COSC-672. The content and requirements within the course do not differ by registration section.)

Course Staff and Office Hours

Nathan Schneider
nathan.schneider@georgetown.edu
Office Hours: Poulton 226 or St. Mary's 342C, by appointment

Textbook

None. All readings will be provided electronically.

Assessments

There will be no major exams. Grades will be based on participation in class, including leading a seminar-style paper discussion, as well as homework assignments such as reading responses and mini-annotation assignments. No programming is required. About half of the grade will be based on a major research project.

Communication

This website and the Canvas platform will be used to host course content. The Canvas discussion forum is the recommended virtual venue for asking and answering course-related questions. The instructor will monitor the forum and post replies from time to time, but cannot promise immediate attention to every question.

The most direct way to contact the instructor is through email.

Instructional Continuity: In the event of a snow day, we will meet virtually via Zoom.

Computing Resources

If you will require a course account on a university Unix server for the project, contact the instructor to request one.

Attendance and Late Policy

In general, students are expected to attend all classes and to complete all assignments on time. Absences may have an adverse effect on grades in a course, up to and including failure.

That being said, we understand that circumstances may arise preventing you from attending class. Please email the instructors ASAP to communicate any expected absences. For example, inform us at the beginning of the semester about planned religious observances or athletic travel.

At the discretion of the instructors, a deadline may be adjusted for a student if there are special circumstances communicated to the instructors well in advance. 11th-hour requests for an extension to an assignment are unlikely to be granted absent truly exceptional circumstances.

Students who miss multiple classes due to prolonged illness should seek medical care and provide documentation of such to the Dean’s Office, which will communicate with the student’s professors. A prolonged absence may necessitate the student’s withdrawal from the course or from the University for the semester.

More information and resources:

Academic Integrity

In this course, you will be asked to participate at times as an individual and at times working in a group.

For homework assignments, you are expected to write code/text and perform analyses yourself unless directed otherwise. I.e., don't copy solutions from other students or share yours with them. But you are encouraged to discuss concepts and implementation stumbling blocks with fellow students, within reason. The online discussion forum and office hours are good opportunities for this.

Part of treating others with respect is giving appropriate credit for ideas and scholarly works (including code). If you consult with other students on an assignment, report this in the work that you turn in. If in you write code using a library or implementation from another source, indicate that as well (minimally by including a URL in a comment). Instructors reserve the right to request an oral explanation of answers.

Course projects are intended to be highly collaborative, and the final project writeup should include a synopsis of who has contributed what. Version control software should be used in development for the final project. (Version control can also be used for homework assignments, provided that you ensure that other students do not have access to your solutions.)

In research writing, it is important to give credit to other research that provides specific foundations to your work, as well as to published work that is closely related. If you discuss ideas/information from a publication, be sure to cite it; if you reuse the specific phrasing of other work, use quotation marks. Knowing when and how to give credit can be tricky at times, so when in doubt, ask!

For more information:

Notice Regarding Sexual Misconduct

Please know that as a faculty member I am committed to supporting survivors of sexual misconduct, including relationship violence, sexual harassment and sexual assault. University policy also requires me to report any disclosures about sexual misconduct to the Title IX Coordinator, whose role is to coordinate the University’s response to sexual misconduct.

Georgetown has a number of fully confidential professional resources who can provide support and assistance to survivors of sexual assault and other forms of sexual misconduct. These resources include:

Jen Schweer, MA, LPC
Associate Director of Health Education Services for Sexual Assault Response and Prevention
(202) 687-0323
jls242@georgetown.edu

Erica Shirley, Trauma Specialist
Counseling and Psychiatric Services (CAPS)
(202) 687-6985
els54@georgetown.edu

More information about campus resources and reporting sexual misconduct can be found at http://sexualassault.georgetown.edu.