Optimization for Large Scale Machine Learning
Logga in till din gruppwebb
Du är inte inloggad på KTH så innehållet är inte anpassat efter dina val.
Gruppwebben som samarbetsyta stängs 1 oktober 2026. Då upphör redigeringsmöjligheterna. Du som administratör kan redan nu välja att exportera gruppens innehåll och radera gruppen.
Mer information och instruktioner hittar du i nyheten: Gruppwebben som samarbetsyta stängs 1 oktober 2026.
Instructor: Prof. S V N Vishwanathan (Purdue University)
News:
- The course begins 9th June, 2014, 10:00-12:00, Room: Q11. The course is for two weeks, comprising nine lectures, in total spanning 18 hours.
- Course duration: 9th June to 19th June. No class on 14th and 15th June.
- Each lecture is about 2 hours. Class time: 10:00 - 12:00.
- Class room address: Q11, Osquldas Väg 6.
Contact persons: Satyam Dwivedi (dwivedi@kth.se) and Saikat Chatterjee (sach@kth.se)
Class time: 10:00-12:00.
About the course
This course is an ACCESS Graduate School course (intensive).
The course is targeted to the graduate students with a research focus on machine learning, or those who are using machine learning in their research. Advanced undergraduates with adequate background can also successfully complete this course.
Prerequisites: Knowledge of linear algebra and probability. Familiarity with a high level programming language (Python preferred). Desirable: some exposure to formal proof techniques (e.g., via a course on real analysis).
- Lecture 1 (Date: 9th June): Introduction to Machine Learning, Few simple yet powerful algorithms.
- Lecture 2 (Date: 10th June): Exponential families of distributions.
- Lecture 3 (Date: 11th June): Density estimation with the exponential family, Maximum a-posteriori vs Bayesian inference.
- Lecture 4 (Date: 12th June): Introduction to convexity, First order and second order methods
- Lecture 5 (Date: 13th June): Conditional models, Logistic regression for binary and multi-class classification.
- Lecture 6 (Date: 16th June): Support vector machines and connections to exponential families, Serial optimization techniques for support vector machines.
- Lecture 7 (Date: 17th June): StreamSVM, Bundle methods.
- Lecture 8 (Date: 18th June): Matrix factorization, Latent Dirichlet allocation (LDA).
- Lecture 9 (Date: 19th June, With Tao together): Distributed and Asynchronous algorithms for matrix factorization and LDA, Questions and clarifications.
General information
Assignment and Materials
(1) Please find the Assignment 1 given by the instructor.
(2) Please find three course materials cm1, cm2 and cm3
(3) Please find two new course materials Optimization1 and Optimization2
(4) Please find a new course material online
