Project in Computer Science: AI
Overview
Students will work in teams to plan, design, implement, test, and document an agentic AI system on a problem or application of their choice in modern AI. Students will learn and practice to build certain parts of a modern AI system, develop a practical application based on AI, design evaluations and improve a project iteratively, assess the project’s ethical, societal, and environmental impacts, and manage a project and work in a team.
Official description in the catalog
4 Units, Discussion, 1 hour; laboratory, 3 hours; research, 3 hours; extra reading, 3 hours. Prerequisite(s): CS 100, CS 111, and CS 170 with grades of “C-” or better; ENGR 180W; 8 additional upper-division units in Computer Science. Covers the planning, design, implementation, testing, and documentation of an artificial intelligence-related system. Incorporates techniques presented in previous related courses. Emphasizes professional and ethical responsibilities; the need to stay current on technology; and its global impact on economics, society, and the environment.Course Details
Instructor: Zhouxing Shi (Email: zhouxing.shi@ucr.edu)
- Discussion: M 5-5:50 PM, UTLF 0215
- Office hours: M 3:30-4:30 PM, WCH 356 (within 351)
TA: Shangjian Yin (Email: syin041@ucr.edu)
- Lab: F 3-5:50 PM, at UTLF 0215
Tentative Schedule
| Week | Date | Discussion (Monday) | Date | Lab (Friday) | Dues |
|---|---|---|---|---|---|
| 1 | 9/28 | Overview | 10/2 | Research and discuss project topics | |
| 2 | 10/5 | Harness basics and tools | 10/9 | Work on project proposal; set up environment | |
| 3 | 10/12 | Context and training | 10/16 | Work on the project | 10/18: Proposal |
| 4 | 10/19 | Midterm presentations | 10/23 | Work on the project | |
| 5 | 10/26 | Midterm presentations | 10/30 | Work on the project | 11/1: Progress report I |
| 6 | 11/2 | Guest Lecture: Justin Cui | 11/6 | Peer review (aim to have an initial version done) | 11/8: Peer review notes |
| 7 | 11/9 | Evaluation and analysis | 11/13 | Work on the project (analysis and improvement) | 11/15: Progress report II |
| 8 | 11/16 | Responsible AI | 11/20 | Work on the project (responsible AI) | |
| 9 | 11/23 | Final presentations | 11/27 | [No Class] Thanksgiving | |
| 10 | 11/30 | Final presentations | 12/4 | Finishing up the project | |
| 11 | 12/11: Final submission |
Grading Overview
Overview:
- Participation (attendance and peer reviews): 20%
- Proposal: 10%
- Midterm presentation: 10%
- Project progress: 10%
- Final presentation: 10%
- Final submission: 40%
Project
Design an AI agent for an application of your choice, with constraints:
- Start from a cheap model and a minimal harness:
- Model: Local models, large open-source models hosted on NRP, cheap commercial models (such as gpt-6-luna), or similar.
- Harness: Start from mini-swe-agent, not Claude or Codex, etc. (You may use a different minimal harness if the suggested one does not apply to your project, e.g., if you will be working on a physical agent.)
- Finish with limited resources:
- If you use paid APIs or GPUs, plan and control your budget carefully, and try to finish the project with a relatively small budget (under $100).
- Include ethical considerations:
- The project should include ethical considerations, implement at least one measure to mitigate risks, and discuss it in the final report. This may be done later in the quarter.
Proposal (10%)
Submit a 1-page write-up to discuss:
- The problem or application you choose and its background
- An initial design of your system
- Resources and budget
- Timeline and team roles
Grading criteria: whether the proposal is reasonably completed on time (due at the end of week 3).
Midterm presentations (10%)
Present in teams on weeks 4 and 5. Each team may use up to 8 minutes, including time for any Q&A. Every team member should present.
The presentation should include:
- Your project topic and a brief overview of what you are building.
- A review of two or three existing relevant projects.
- Suggested content: what they did, what features they have, their overall design, their limitations, and comparison with what you are going to do.
- If time allows and information is available, you may mention selected implementation details that are relevant to your project.
- This is for you to better understand the landscape of existing work. Your project is supposed to be independent from them, and you don’t have to build your project based on them, but you may borrow some techniques from those existing projects.
Slides are recommended, though not required (you may use the whiteboard if you like).
Please upload your midterm presentation slides (or notes if you don’t use slides) by the end of your presentation day to the assignment on Canvas.
Project progress (10%)
Submit a brief (~0.5 page) progress report on weeks 5 and 7. Each student needs to write an individual report and the report should reflect your individual work within the team.
Peer review
A peer review is planned for the lab in week 6. Each team will match with another team. The two teams will discuss each other’s project so far and provide feedback. Each team member should contribute to the discussion. Submit a brief discussion summary at the end (per team). This counts towards the participation grade.
Final presentation (10%)
In-class demo in week 9 or 10 (10%). The project does not have to be complete yet.
Final submission (40%)
Submit a final demo (short video), report, and code, due at the end of week 11.
Conduct and Academic Integrity
Collaboration. All team members should contribute roughly equally. Document each person’s contributions in the final report.
Originality. Your project must include substantial work created by your team. Cite any existing materials (code, models, datasets, etc.) you reuse.
AI use. You may use AI tools to assist with coding. Review and understand AI-generated content, and document your usage.
See https://conduct.ucr.edu for details. Here at UCR we are committed to upholding and promoting the values of the Tartan Soul: Integrity, Accountability, Excellence, and Respect. As a student in this class, it is your responsibility to act in accordance with these values by completing all assignments in the manner described, and by informing the instructor of suspected acts of academic misconduct by your peers. By so doing, you will not only affirm your own integrity, but also the integrity of the intellectual work of this University, and the degree which it represents. Should you choose to commit academic misconduct in this class, you will be held accountable according to the policies set forth by the University, and will incur appropriate consequences both in this class and from Student Conduct and Academic Integrity Programs. For more information regarding University policy and its enforcement, please visit: conduct.ucr.edu.