Computational Models of Discourse and Dialogue: Sentiment, Sarcasm and Mood in Social Media

CMPM 245, Winter 2015

Computational Models of Discourse and Dialogue

Topic: Sentiment, Sarcasm and Mood in Social Media

CMPM 245 (formerly CMPS 245) is an advanced seminar in Natural Language Processing with a focus on current research and methods in the computational modeling of discourse and dialogue phenomena in social media, online dialogue and personal (first person) narratives. Processing naturally occurring discourse and dialogue in many cases relies on some representation of the context, the context of the dialogue in terms of what was said, the social relationship of the conversants, or other aspects of the situational context.

In Winter 2016 the class topical focus is Sentiment, Sarcasm and Mood in social media. We will draw on a number of different theories of these language phenomena as well use a range of different datasets, read computational papers on modeling these phenomena, and develop a project based on a dataset and a computational or corpus-based modeling approach.

The class is a combination of readings, homeworks, a midterm, presentations and final project. The final project report is due on the scheduled day of the final exam. Class presentations of final projects will be during the last week of class.

Any student who thinks s/he is in need of accommodation, based on the impact of a disability, should contact me privately to submit their Accommodation Authorization and discuss specific needs, preferably within the first two weeks of the quarter. Please contact the Disability Resource Center at 831-459-2089 in room 125 Hahn Student Services or by e-mail at drc@ucsc.edu to coordinate those accommodations.

Course Format and Grading

10%: Short Written essays on readings, and participation in class discussion on the basis of these reflections.
In-class Presentations (2 or more per person). These must be semi-formal. These should be 25 minute presentations for conference papers and 35 minutes for journal papers. They should be given using a projector with powerpoint or other similar software. They may include a handout to accompany the lecture if desired. (20%)
Written Homeworks: The homework assignments will familiarize you with tools that will be useful for your project. They will not be graded except that material from them will appear on the midterm.
Midterm: 30%. The Midterm will be multiple choice and will be based on homework and readings.
Final Project (40%): Divided into Proposal (10%) which is due Thurs Feb 5th, Final Presentation and Written Report (30%) due Wed Mar 18th 11AM. Project reports must be written using ACL format for conference submission, and should be written in the style of an ACL paper.
Homework Delivery: Turn it in on the course assignments page using MS Word or PDF formats. Okay to scan.
Please refer to the university policy on academic integrity.