🌏 中文版
Video status: Checked: no corresponding recording link listed on the public official page. Source details
CMU 07-380 AI & ML II Fall 2026 follows 07-280 Fall 2026 as the second course in the new sequence, first offered (New in Fall 2026). It is not a reshuffle of 15-281/10-315 but an extension that lifts the search/ML/RL/probability spine of 07-280 into deeper reasoning, optimization, graphical models and generative systems.
Why 07-380 exists
07-280 Fall 2026 already covers 24 lectures + HW0-11 + weekly Checkpoints from heuristic search to AlphaZero. The FAQ states 07-280/07-380 will be offered every semester, 07-380 first being offered in Fall 2026, jointly replacing 15-281 + 10-315. Prerequisites are 07-280 + probability (36-225 etc.) with C or above; 10-301 holders need 10-301 + 16-350 reviewed via bsai@cs.cmu.edu (see Prerequisites).
Grading differs from 07-280: 55% quizzes (6, drop lowest) + 25% HW + 10% project + 5% checkpoints + 5% participation, no final exam, capped by the Final Project (Dec 7-15).
26-lecture arc
Per the Fall 2026 Schedule (subject to change):
| Block | Lectures | Topics |
|---|---|---|
| Reasoning | Lec02-04 | Logical Agents / SAT-DPLL / Planning PDDL / Motion RRT |
| Optimization | Lec05-07 | LP / ILP / PCA + LoRA |
| Reasoning under uncertainty | Lec08-13 | MAP / Naive Bayes / Bayes Nets / GMM-EM / Sampling-Gibbs / HMM-Particle Filter |
| Acting & generative | Lec14-18 | RL Policy Gradient→RLHF / DL Optimization / VAE / Diffusion / Multimodal VLM/VLA/JEPA |
| Systems & theory | Lec19-24 | Parallel (data/model/pipeline/MoE) / Scaling Laws / Recommender / Ensemble / PAC-VC / Game Theory |
| Practice | Lec25-26 | ML in Practice / AI Ethics + Quiz 6 |
How public is it
As of the course site on 2026-09-29 (the site says subject to change):
| Item | Anonymous | Gap |
|---|---|---|
| Syllabus / Schedule / Policies | fully public | — |
| Lecture slides | Lec1–8 each have pptx/pdf plus inked versions; one shared Lec9-10_Probabilistic_Generative_Models deck | the Lec10 Bayes Nets deck is not up (the guessed URL returns 404); Lec11–26 list only topics and readings |
| Pre-reading notes | PR1 Prop Logic, PR2 Planning, PR3 LP, PR4 PCA, PR5 MAP, PR6 Bayes Nets, plus Notes_Generative_Models | PR7–13 have dates only |
| Recitation | handouts and solutions for Rec1, Rec2, Rec3-4, Rec5 | no handouts from Rec6 on |
| HW | HW1 (logic_plan programming), HW2 (planning programming + written PDF/LaTeX template), HW3 (optimization programming + written); starter zips and local autograders download freely | HW4–7 have tentative dates only; Gradescope online questions are campus-only |
| Quiz / Project | quiz dates posted | quiz questions are not public; project groups, proposal and submission are still TBA |
| Platforms | described | Canvas/Gradescope/Piazza campus-only; no recording links on the site |
On this site's A0–A3 scale: the course as a whole is A2 (in progress). The released Lec1–9 block has slides, notes, recitation solutions and assignment starters, which is A3: you can self-study through HW3.
This series, lecture by lecture
The 14 guides below follow a learning order (each assignment comes after the last lecture it covers), all based on course materials as of 2026-09-29:
| order | Post | Scope |
|---|---|---|
| 1 | Lecture 1: Introduction | what II adds after 07-280, grading structure |
| 2 | Lecture 2: Logical Agents | model checking, DPLL, forward chaining |
| 3 | HW1: Logic and the Hybrid Wumpus Agent | SAT calls to judge safe squares, building the hybrid agent |
| 4 | Lecture 3: Classical Planning | PDDL, state-space search, relaxation heuristics |
| 5 | Lecture 4: Motion Planning (RRT) | sampling in continuous space, RRT and RRT* |
| 6 | HW2: Classical and Motion Planning | robot-cook PDDL, RRT*, graphing LPs |
| 7 | Lecture 5: Linear Programming | why the optimum sits at a vertex |
| 8 | Lecture 6: Integer Programming | LP relaxation and branch and bound |
| 9 | Lecture 7: PCA and Low Rank (LoRA) | reconstruction error, projected variance, SVD |
| 10 | Lecture 8: MAP | priors as regularization |
| 11 | HW3: Optimization | your own LP solver, branch and bound, PCA and MAP on paper |
| 12 | Lecture 9: Generative Models | Naive Bayes and GDA |
| 13 | Lecture 10: Bayes Nets (pre-reading edition) | PR6 notes only; slides not up |
| 14 | Stage recap: certainty and optimization → uncertainty | the first ten lectures and HW1–3 in one view |
Not released yet, to be added later
Lec10 Bayes Netsslides: order 13 will be updated from them once they are posted.Lec11onward (Approximate Inference, HMM/Particle Filter, GMM/EM, Policy Gradient→RLHF, VAE, Diffusion, the systems and theory lectures): the site lists only topics and readings.HW4–7,PR7–13, handouts fromRec6on, and the Final Project spec.- Campus-only, so this series will not cover them: quiz questions, Canvas checkpoints, Gradescope online questions.
Bridge from 07-280
07-280 I (24: search/CSP → ML/NN→CNN/Transformer → RL/MCTS)
↓
07-380 II (26: logic/planning → optimization → graphical/HMM → RLHF/generative → systems/theory/ethics)07-280 builds AlexNet/GPT-2/AlphaZero; 07-380 formalizes reasoning and optimization, exact/approximate inference on graphical models, and scalable generative systems. Off-campus, finish 07-280 as the spine, then backfill formal depth via LP/PCA/MAP/EM/Bayes Nets/HMM/VAE/Diffusion.
Tonight's starter
- Diff
07-380 Lec01(pdf) against07-280 Lec01to markI→IIoverlap and new units (Planning/PDDL, LoRA, PAC-VC). - If logic or optimization is weak, read
Prop Logic.pdfandBoyd Ch.2/4before jumping toDiffusion. - Start from order 1 above; after each block, download the matching HW starter and run its autograder locally.
Update log
- 2026-10-10: Added explicit video status and checked recording sources and access notes.
- 2026-10-10: Rechecked video status. Re-verified the live official course pages and public video sources; no public recording for this lecture was found, so the status stands.
- 2026-09-29: Rewrote "How public is it" for the 2026-09-29 course site (Lec1–9, PR1–6, Rec1–5 and HW1–3 are public) and changed the rating to "A2 overall, A3 for the Lec1–9 block"; added the 14-post series list and the pending list; changed starter step 3 to begin at order 1.
Course video sources
The official Fall 2026 schedule and assignment list have been checked: public resources include slides, pre-readings, demonstrations and assignments, but no public recording link for the corresponding lectures. This article is therefore a materials-based guide with no corresponding lecture player. This observation concerns the public official page and does not establish whether internal recordings exist.
Official sources:
Checked on 2026-10-10.
References
Loading...