Spectral Clustering (2024 fall) Final date for presentation of papers or real life data analysis is 28 November, in class. Those, who solve the theoretical exercises, the deadline for turning in the solutions is 5 December. Those, who will have earned the signature, can write a final test on 5 December (exam topics), the result of which can be improved by oral exam in the exam period.

---

• Exam topics
• Course requirements
• Topic description
• Midterm assignments
• Papers for presentation (it is indicated if a paper has already been selected by someone)
• Material of Lesson 1 (Quadratic placement and Laplacian matrices, hypergraphs can be skipped)
• Material of Lesson 2 (Normalized Laplacian and modularity matrices, RKHS, rectangular arrays can be skipped)
• Material of Lesson 3 (Normalized cuts and the isoperimetric number, bicuts of rectangular arrays can be skipped)
• Material of Lesson 4 (Random block matrices, discrepancy; rectangular arrays can be skipped)
• Material of Lesson 5 (Graph convergence, testable graph parameters; only the notion of graph convergence and graphon is neede)
• Material of Lesson 6 (Generalized random and quasirandom graphs and their properties)
• Non-backtracking matrix for simple graphs, sparse stochastic block model, see the ArXiv paper 2307.16502v6 on my homepage under Publications.
• Material of Lesson 7 (Parameter estimation via the EM algorithm in probabilistic block models)
(does not belong to the material in this semester)
• Basics: Real Matrices
• Basics: Random Matrices
• We follow the book (some parts will be skipped):
  Spectral Clustering and Biclustering. Learning Large Graphs and Contingency Tables, 2013, Wiley.