Lecture notes are taken by students. The scribe is expected to verify the correctness of all proofs, to fill in any obvious gaps in the lecture, and to add figures and illustrative images as needed. Please use the provided .sty file, and follow the style and notation conventions from the sample document.

Please sign-up to scribe for a lecture using this sign-up spreadsheet.

Most of these notes have been posted as submitted, without careful checking by the instructors; feel free to contact us with any questions or potential issues with these notes.

Title File
1 Introduction to Online Learning PDF
2 Online Convex Optimization and Follow the Leader PDF
3 Follow-The-Regularized-Leader PDF
4 Convexity and Online Gradient Descent PDF
5 FTRL with Arbitrary Strongly Convex Regularization and Experts PDF
6 Online Learning with Expert Advice PDF
7 Exponentiated Gradient and Bandits PDF
8 The Multi-Armed Bandit Problem PDF
9 Tight EXP3 analysis PDF
10 Bandits with Expert Advice: From EXP3 to EXP4 PDF
11 Analyzing Adaptive Algorithms I PDF
12 Analyzing Adaptive Algorithms II PDF
13 Adaptive Algorithms III PDF
14 Stochastic Experts and Bandits PDF
15 Stochastic bandits: Explore-First and UCB PDF
16 Bandit Online Convex Optimization PDF
17 Combinatorial Bandits PDF
18 Minimax Optimal Algorithms PDF
19 Information-Theoretic Lower Bounds PDF
20 CTR Predictions and Literature References
(References) (CTR Prediction Talk)

Latex Template

Lecture notes should be written in Latex using the cse599s14sp.sty style file and compiled into pdf using pdflatex. Here is an example of a latex file that uses the style file, and here is the resulting pdf document.