CSCI 680 · Affective Computing Overview Schedule Work Policies

Six componentsNo final exam30% is your project

Work & grading

Most of the grade rewards two things: showing up having actually read the paper, and finishing a piece of original research.

Grade composition
ComponentShareWeight
Final project — report and presentation
30%
Paper presentations — one or two over the semester
25%
Reading responses
15%
Seminar participation
15%
Midterm progress report
10%
Project proposal
5%

Reading responses

Due by 9:00 a.m. on the day of each class you are not presenting. Roughly 250–400 words. Graded credit / no-credit on evident engagement.

Your two lowest scores are dropped, no questions asked. They exist to absorb ordinary life — a conference deadline, a bad week, a flight. Use them without explaining.

Late responses are not accepted. The whole point is that you read the paper before we discuss it.

Each response covers three things

  • A one-paragraph summary of the paper's central claim.
  • One substantive critique — a methodological gap, an unconvincing experiment, an unexamined assumption, or a claim the evidence does not support.
  • One question you want the seminar to discuss.

Paper presentations

You present once or twice over the semester. Sign-ups happen at the first meeting on August 27.

Plan for 35–40 minutes covering the problem, the method, the experiments, and the weaknesses — then facilitate about 30 minutes of discussion. Facilitating is part of the grade: come with questions that the room can't answer from the abstract alone.

The weaknesses section is not optional

Every paper on the schedule was accepted somewhere good, and every one of them oversells something. Finding that thing is most of what this course teaches.

Final project

An original research project, done individually or in pairs, worth 30% of the grade — plus 15% across the proposal and progress report that lead into it.

Acceptable project types

  • A novel method with empirical evaluation.
  • A careful reproduction and stress-test of a published result.
  • A rigorous empirical study of an open question — for example, how a particular class of model degrades under a specific kind of shift.

Not acceptable

Pure literature surveys. If the project doesn't run something and measure something, it isn't a project for this course.

If you're collecting new data

Any project collecting new data from human participants needs IRB review before collection begins. Build that timeline into your proposal — see research ethics.

  • Friday, September 18 Project proposal — 2 pages. Worth 5%.
  • Friday, October 30 Midterm progress report — 4 pages. Worth 10%.
  • December 1 and 3 In-class presentations — sessions I and II.
  • Friday, December 11 Final report — 8 pages, conference format. No final exam.

Late project work

Project deliverables may be submitted up to 48 hours late with a 10% penalty. Beyond that, only by prior arrangement. Reading responses are never accepted late.

Everything with a date on it