EE290S, Fall 2008
Network Information Flow
Tues. and Thurs.: 3:30 pm - 5:30 pm
- Course Description:
Claude Shannon invented information theory in 1948 and it forms the basis for the design of all modern day communication systems. His original theory was primarily point-to-point, quantifying information and studying how fast it can flow across an isolated noisy communication channel. Until recently, there has been only limited success in extending the theory to a network of interacting nodes. Progress has been made in the past decade, driven both by engineering interest in wireless networks as well as conceptual advances such as network coding, which shows the surprising result that even in noiseless wireline networks, coding across packets can improve throughput beyond pure store-and-forward routing.
This course is a fresh perspective on the state-of-the-art of the field. Our strategy is to use analytically simpler discrete deterministic network models to capture the interaction between nodes and as an approximation to more complex noisy network models. Starting with the classical wireline network and progressing to more complex models for wireless networks, we hope to answer questions such as:
What is the optimal way for relay nodes to cooperate and send information?
How can information flow over a network be quantified and visualized?
What is the optimal way to deal with interference between two or more competing flows of information?
Because of our emphasis on the use of deterministic models, we hope to make this course accessible to computer science and electrical engineering students alike.
- Course Requirements:
All students will be required to do an in-class presentation of a paper or two (depending on the time) and a final project.
- Raymond Yeung, Information Theory and Network Coding, ch. 17 - 21 (available in pdf)
- Course Outline:
Course outline available here.
Professor David Tse
257 Cory Hall
Phone: (510) 642-5807
Office Hours: Mon. and Wed., 9:00-10:00 am
Professor Anant Sahai
Phone: (510) 642-9719
Office Hours: Mon. 4:00-5:00 pm, and Tues. 2:30-3:30 pm in 258 Cory
Course Administrative Assistant:
Phone: (510) 643-4976
- Sept. 12, 2008
Future Scribes, when sending your notes, please zip and send the whole directory for your scribed notes.
- Aug. 28, 2008
Students should add the course and use bSpace to access more information.
Welcome to EE290S
Back to top
- Aug. 28, 2008 - Lecture # 1:
- Sept. 2, 2008 - Lecture # 2:
Ford-Fulkerson Max Flow Min Cut Theorem
- Sept. 4, 2008 - Lecture # 3:
Linear Network Coding
- Sept. 9, 2008 - Lecture # 4:
Existence of Network Codes
- Sept. 11, 2008 - Lecture # 4.5:
Jaggi et.al algorithm for Direct Acyclic Graphs
- Sept. 16, 2008 - Lecture # 5:
- Sept. 18, 2008 - Lecture # 6:
Cyclic Networks with LTI Codes
- Sept. 23, 2008 - Lecture # 7:
- Sept. 25, 2008 - Lecture # 8:
Most General Channel Model
- Sept. 30, 2008 - Lecture # 9:
Recap: Binary Expansion Model
- Oct. 2, 2008 - Lecture # 11:
- Oct. 7, 2008 - Lecture # 11-2:
Extending the min-cut theorem to general linear deterministic
network models that are not necessarily layered or acyclic.
- Oct. 9, 2008 - Lecture # 12:
Linear Deterministic Models of Relays
- Oct. 14, 2008 - Lecture # 14:
Natural Extensions of the Single Source Multiple
- Oct. 16, 2008 - Lecture # 15:
Natural Extensions of the Single Source Multiple Destinations Case (cont.)
- Oct. 21 & 23, 2008 - Lecture # 16:
Finishing up on earlier lectures
- Oct. 30, 2008 - Lecture # 18:
- Nov. 4, 2008 - Lecture # 19:
The Converse for the Gaussian Broadcast Channel
- Nov. 6, 2008 - Lecture # 15:
The Relay Channel
- Nov. 11, 2008 - Lecture # 21:
Gaussian relay channels
- Nov. 18, 2008 - Lecture # 23:
General Layered Network
- Nov. 20, 2008 - Lecture # 24:
Review of Previous Lecture
- Nov. 25, 2008 - Lecture # 25:
- Dec. 2, 2008 - Lecture # 26:
Linear Deterministic Interference Channel
- Dec. 4, 2008 - Lecture # 27:
Recap - Linear Deterministic Interference Channel
Back to top
Back to top