CSCI-B659 - Spring 2016
Stochastic optimization for machine learning
When: Monday 2:30 pm-5:00pm
Where: AC C112
Office: LH 401E
Office Hours: By appointment
Office: LH 406
Office Hours: Wednesday 4-6 p.m.
Many machine learning algorithms are based on stochastic optimization. In this course, we will focus on learning from streams of data, and explore efficient incremental algorithms and their theoretical guarantees. We will look at time series data, online learning problems and reinforcement learning. Much of the course will be about prediction on streams, but we will also look at control, where an agent learns to make decisions. The main purpose of this course is to be an intensive development course for research in machine learning. In addition to modern topics about incremental estimation in machine learning, there will be a focus on developing your knowledge of how to do research in machine learning (including top conferences, how to publish, etc.) and how to read and understand machine learning papers.
CSCI-C 555 (Machine Learning), or permission from the instructor. The course will require some basic machine learning knowledge, to better grasp these more advanced topics, and the ability to program the algorithms in some language (e.g., python, C). A good course to take in parallel is the Reinforcement Learning course, B-659.
Reading list of machine learning papers provided.
20%: Two assignments, each 10%
20%: Readings, discussion and short presentations
60%: Research project
The class meets once a week, for three hours. Out of class, you will be reading and understanding papers, to prepare to discuss them in class. You will also be required to present parts of papers. The focus of the course is on doing research in machine learning, with a strong focus on your research projects (which I encourage to do in pairs). I will provide several research projects that fit very well with the topic of the class, but you are also free to work on a project of your design. In class, there will be breaks to make the three hours more manageable, but the three hours gives us time to properly discuss a paper, and whatever topics we are covering at that time.
Late Policy and Academic Honesty
All assignments and exams are individual, except when collaboration is explicitly allowed. All the sources used for problem solution must be acknowledged, e.g. web sites, books, research papers, personal communication with people, etc. Academic honesty is taken seriously; for detailed information see Indiana University Code of Student Rights, Responsibilities, and Conduct.
The syllabus page shows a table-oriented view of the course schedule, and the basics of course grading. You can add any other comments, notes, or thoughts you have about the course structure, course policies or anything else.
To add some comments, click the "Edit" link at the top.