Materials on this page are tentative until classes begin!
Course Assessment Measures
At the 4000 level, the grade will be based on reading and participation (worth 20%), a final project (worth 30%), and the average homework grade (worth 50%).
The homeworks and course project can be done in groups of at least two and up to three; all members of the group will get the same grade on the assignments. In general, each homework will involve a reasonable amount of programming, either from scratch or by integrating existing software tools/libraries and Google Colab notebooks. Homework is due at the start of class on the date indicated. Late submissions will not be accepted since the grading process also involves your assessment of others’ work.
Tentatively, the assignments will address:
- Dataset selection/curation, variational autoencoders, and project proposal (due 9/21, 10% of grade)
- GANs and diffusion models (due 10/5, 10% of grade)
- Large language models (due 10/30, 10% of grade)
- Text to image and video (due 11/16, 10% of grade)
- Neural rendering fields (NeRF) (due 11/30, 10% of grade)
The 6000 level of the class has two additional requirements. First, each homework will include an additional problem involving a critical 2-page review (not just a summary) of 2 recent papers from CVPR, SIGGRAPH, NeurIPS, etc. that pertains to the material for that homework. Second, the 6000-level students will also be expected to incorporate more research-level/hand-written code into their assignments and final projects.
Each assignment will include a peer evaluation component distributed via Google Forms; the final grade for each assignment will be a weighted average of 90% the instructor’s assessment and 10% the average peer assessment.
The participation grade (20% of the course grade) will be made up of 50% instructor assessment of student engagement/contribution, 25% within-group peer assessments, and 25% frequency of peer assessment submissions. Within-group assessments will be monitored per assignment to assess any imbalances that should affect the participation grade.
The final project of the students’ own design should be a creative product that leverages at least three of the concepts from the class. Students will provide one progress report during the semester, and a final written report and source code on the last day of class. The overall project grade will be computed as 10% proposal, 20% progress report, 35% final report, and 35% project evaluation. All projects will be presented and discussed in the last week of class.
The details of the grading policy are provided below.
Grading policy for 4964
- Homework assignments 50% (10% each)
- Class participation 20%
- Final project 30% (broken down as above)
Grading policy for 6964
- Homework assignments 50% (10% each)
- Class participation 15%
- Final project 25% (broken down as above)
- Reaction reports to research papers 10%
Course Policies:
Attendance is expected in every class period, unless previously discussed with the instructor, and if necessary, officially documented by the Student Experience office (4th floor, Academy Hall). We will cover a lot of ground in this course, so attendance is important (especially during the activity days). Note that a percentage of your course grade also depends on your participation and engagement during class.
You are expected to approach the instructor with any issue that may affect your performance in class ahead of time. This includes absence from important class meetings, late assignments, inability to perform an assigned task, the need for extra time on assignments, etc. You should be prepared to provide sufficient proof of any circumstances based on which you are making a special request as outlined in the Rensselaer Handbook of Student Rights and Responsibilities.
Grade appeals on an assignment should be made in Gradescope within 72 hours of its return to the class. Letter grades will not be assigned until the end of the class, after the final project report has been graded. Any letter grade assignment posted before the end of the class should be regarded as tentative and subject to change.
Students acknowledge the use of various third-party software and tools (Piazza, Gradescope, Google Forms, Discord, Colab, OpenAI, etc.) that may require them to disclose personal information outside of the RPI network
Academic Integrity
Student-teacher relationships are built on trust. For example, students must trust that teachers have made appropriate decisions about the structure and content of the courses they teach, and teachers must trust that the assignments that students turn in are their own. Acts that violate this trust undermine the educational process. The Rensselaer Handbook of Student Rights and Responsibilities and The Graduate Student Supplement define various forms of Academic Dishonesty and you should make yourself familiar with these. In this class, all assignments that are turned in for a grade must represent the student’s own work. In cases where help was received, or teamwork was allowed, a notation on the assignment should indicate your collaboration.
Collaboration between teammates is allowed and encouraged on the homeworks and project, but collaboration with others within or outside RPI is prohibited. Submission of any assignment that is in violation of the collaboration policies will result in a penalty of an F in the class, and may be subject to further disciplinary action.
Academic Accommodations
Rensselaer Polytechnic Institute is committed to providing equal access to our educational programs and services for students with disabilities. If you anticipate or experience academic barriers due to a disability, please contact the Office of Disability Services for Students (DSS) (dss@rpi.edu; 518-276-8197) to establish reasonable accommodations. Once you have been approved for accommodations, please provide your Faculty Memorandum (a letter provided to students by DSS) to all faculty members for this course. Please provide this at the very beginning of the semester.