David Steurer: Katalogdaten im Frühjahrssemester 2021

NameHerr Prof. Dr. David Steurer
LehrgebietTheoretische Informatik
Adresse
Professur Theoretische Informatik
ETH Zürich, OAT Z 22.2
Andreasstrasse 5
8092 Zürich
SWITZERLAND
E-Maildavid.steurer@inf.ethz.ch
URLhttps://www.dsteurer.org
DepartementInformatik
BeziehungAusserordentlicher Professor

NummerTitelECTSUmfangDozierende
252-4202-00LSeminar in Theoretical Computer Science Information 2 KP2SE. Welzl, B. Gärtner, M. Ghaffari, M. Hoffmann, J. Lengler, D. Steurer, B. Sudakov
KurzbeschreibungPresentation of recent publications in theoretical computer science, including results by diploma, masters and doctoral candidates.
LernzielTo get an overview of current research in the areas covered by the involved research groups. To present results from the literature.
Voraussetzungen / BesonderesThis seminar takes place as part of the joint research seminar of several theory groups. Intended participation is for students with excellent performance only. Formal restriction is: prior successful participation in a master level seminar in theoretical computer science.
252-4225-00LPresenting Theoretical Computer Science Belegung eingeschränkt - Details anzeigen
Number of participants limited to 24.

The deadline for deregistering expires at the end of the second week of the semester. Students who are still registered after that date, but do not attend the seminar, will officially fail the seminar.
2 KP2SB. Gärtner, M. Ghaffari, R. Kyng, D. Steurer, E. Welzl
KurzbeschreibungStudents present current or classical results from theoretical computer science.
LernzielStudents learn to read, understand and present results from theoretical computer science. The main focus and deliverable is a good presentation of 45 minutes that can easily be followed and understood by the audience.
InhaltStudents present current or classical results from theoretical computer science.
Voraussetzungen / BesonderesThe seminar takes place as a block seminar on two Saturdays in April and/or May. Each presentation is jointly prepared and given by two students (procedure according to the seminar's Moodle page).
All students must attend all presentations. Participation requires successful completion of the first year, or instructor approval.
261-5110-00LOptimization for Data Science Information 10 KP3V + 2U + 4AB. Gärtner, D. Steurer, N. He
KurzbeschreibungThis course provides an in-depth theoretical treatment of optimization methods that are particularly relevant in data science.
LernzielUnderstanding the theoretical guarantees (and their limits) of relevant optimization methods used in data science. Learning general paradigms to deal with optimization problems arising in data science.
InhaltThis course provides an in-depth theoretical treatment of optimization methods that are particularly relevant in machine learning and data science.

In the first part of the course, we will first give a brief introduction to convex optimization, with some basic motivating examples from machine learning. Then we will analyse classical and more recent first and second order methods for convex optimization: gradient descent, Nesterov's accelerated method, proximal and splitting algorithms, subgradient descent, stochastic gradient descent, variance-reduced methods, Newton's method, and Quasi-Newton methods. The emphasis will be on analysis techniques that occur repeatedly in convergence analyses for various classes of convex functions. We will also discuss some classical and recent theoretical results for nonconvex optimization.

In the second part, we discuss convex programming relaxations as a powerful and versatile paradigm for designing efficient algorithms to solve computational problems arising in data science. We will learn about this paradigm and develop a unified perspective on it through the lens of the sum-of-squares semidefinite programming hierarchy. As applications, we are discussing non-negative matrix factorization, compressed sensing and sparse linear regression, matrix completion and phase retrieval, as well as robust estimation.
Voraussetzungen / BesonderesAs background, we require material taught in the course "252-0209-00L Algorithms, Probability, and Computing". It is not necessary that participants have actually taken the course, but they should be prepared to catch up if necessary.