Schedule
Tentative and updated as the quarter goes. Slides open in presentation mode (press F for fullscreen, S for speaker view, E for print/PDF mode).
| # | Date | Topic | Materials | Reading | Recommended Exercises |
|---|---|---|---|---|---|
| 1 | 9/29 | Why eigenvectors? A tour of spectral methods | Slides · demos: PageRank, Fiedler vector, Heat diffusion | LA Review I | Perron–Frobenius · ℓ₁ contraction |
| 2 | 10/1 | Graphs as Differential Operators. Random Walks, the Heat Kernel, Measures vs. Densities. | Slides · demo: Split and aggregate | Lecture 2 notes §1–2 | Computing with Laplacians, Ex. 1–7 |
| 3 | 10/6 | Random Walks and their Convergence: Mixing Times, the Spectral Gap, Canonical Paths | Slides | Lecture 2 notes §3–6 | Computing with Laplacians, Ex. 8–9 |
| 4 | 10/8 | Spectral Clustering and Eigenvector Perturbation | Davis–Kahan | ||
| 5 | 10/13 | Conductance and Cheeger’s Inequality | |||
| 6 | 10/15 | Higher-order and k-way Cheeger’s Inequalities | |||
| 7 | 10/20 | Electrical Flows and Fast Laplacian Solvers | |||
| 8 | 10/22 | Semidefinite Programming and MMWUs | |||
| LA Review II |