6.454 Graduate Seminar in Area I - Fall 2008
W 1:15-2:15pm, 32-D677
Contact: Devavrat Shah (firstname.lastname@example.org, x34670).
Student organizers: Jinwoo Shin (email@example.com), R Shreevatsa (firstname.lastname@example.org)
This is a student run graduate seminar course exploring advances in Area I topics including communication, control, networks, signal processing, and optimization. Relationships to related fields, including artificial intelligence, computer science, machine learning and operations research are also considered.
The class meets once a week starting September 3rd. An initial set of topics (stated below) will be discussed in the first meeting. Students are encouraged to bring a topic of their choice to the first meeting. During this organizational meeting, participants will be encouraged to select topics (and related papers/book chapters) and divide themselves into small groups (one or two students per group). Each subsequent week, a student group will do a recitation-like presentation of their chosen topic. Prior to the presentation, the student group is required to review the literature in sufficient detail, with faculty assistance as required, and produce a written summary suitable for distribution.
Tentative list of topics (Look at the Topics page for details):
- Markov Decision Processes and Bandit Problems
- Optimization relaxation methods.
- Learning the "un-learnable": The Good-Turing estimator, NetFlix contest, Network tomography, etc.
- Graph convergence, algorithms and property testing : local weak convergence and dense graph convergence.
- Information-theoretic security and cryptography.
This course was originally proposed in 2000. It has been developed and maintained since by students in Area I, mainly from the Laboratory for Information and Decision Systems (LIDS) and the Research Laboratory of Electronics (RLE), with the guidance of several members of the faculty.
Websites and presentation materials from previous terms are kept available online: