The schedule is subject to change: this website is under construction; please check back frequently.
Meetings: fourteen Mondays, 5:00–8:20 PM, from August 24 to November 30. The university is closed on September 7 (Labor Day). Section 30122 meets in SGM 124; section 30359 attends through DEN@Viterbi. The August 31 meeting is a recorded lecture for both sections: there is no live session and no meeting in SGM 124 that day.
Discussion: Piazza
Quiz policy: As discussed in the first class, quizzes are dynamic: we may give a quiz if in-person attendance falls below 50%. Quizzes have no fixed dates. For any quiz not administered, its grading weight is transferred to the project final report and presentation. See Piazza @101.
Assignment: AI-Assisted Image Classification is available early (password shared in the Piazza announcement). The deadline remains October 26, 23:59 Pacific Time. Post all assignment questions on Piazza as private posts to instructors. Do not post assignment code, results, or report excerpts publicly; staff will share general clarifications.
| # | Date | Topic | Assessment and deadlines |
|---|---|---|---|
| 1 | Aug 24 |
Course intro; ML and DL framing; agent security and trustworthy AI; linear models
Slides |
|
| 2 | Aug 31 |
Classical ML: decision trees, kNN, clustering; model metrics
Slides Recorded lecture, no live session. There is no meeting in SGM 124 that day; watch the recording on your own time. |
|
| — | Sep 7 | Labor Day, no class | |
| 3 | Sep 14 | Neural networks: forward propagation, backpropagation, gradient descent, activation functions, and training stability Slides (Block 3) | Pre-proposal due |
| 4 | Sep 21 | Convolutional neural networks; visual representations and CLIP; NVIDIA guest lecture Slides (Block 4) | |
| 5 | Sep 28 | RNNs and LSTMs; attention for image captioning; Transformer basics and parallel training Slides (Block 5) | |
| 6 | Oct 5 | Graph representations; graph neural networks and message passing; connections to LLMs Slides (Block 6) | |
| 7 | Oct 12 | Training dynamics; transfer learning Slides (Block 7) | Assignment already available (released early; one assignment this semester) |
| 8 | Oct 19 | Generative AI: GAN and VAE; introduction to diffusion and flow matching Slides (Block 8) | Exam scope posted |
| 9 | Oct 26 | Guest lecture: Ruolin Li, AI for intelligent transportation (tentative topic); attention in depth; decoding, KV cache, and mixture of experts; midterm review Slides (Block 9) | Midterm Review | Assignment due at 23:59 Pacific Time |
| 10 | Nov 2 | Midterm | Midterm |
| 11 | Nov 9 | Self-supervised learning and LLMs; out-of-distribution and anomaly detection Slides (Block 10) | Project midterm report due |
| 12 | Nov 16 | Guest lecture: Ruishan Liu, AI for biomedicine and healthcare; scaling laws; reinforcement learning | |
| 13 | Nov 23 | Guest lecture: Mengyuan Li (5:00-6:00pm); automated ML; project clinic | |
| 14 | Nov 30 | Project poster session | |
| — | Dec 7 | No final exam | Final report due |
Midterm: the exam covers everything taught up to and including the October 26 lecture. A narrower list of what is and is not tested is posted here and on Piazza by October 19; that posted scope only narrows the material above, never adds to it.
Final presentation: the last meeting is a poster and demo session rather than sequential talks. All groups present in parallel, and every group member answers questions individually.