Information Theory in Computer Science (Harvard CS 229r)

General Info:

• Lecturer: Madhu Sudan;  MD 339; email: first name at cs dot harvard dot edu; Office Hours: Thursday 2:30-4pm
  
• TF: TBD
  
• Time and Location: TuTh 1-2:30 in MD 119.

Announcements:

• Course announcement
• Grading Policy
• Scribe availability (Sample files for scribes: sample.tex; Uses: preamble.tex and fullpage.sty)
• Piazza site. (Please make sure you've joined this site to ensure you're getting all announcements.)

Lecture Notes:

• Lecture 1: Introduction. Shearer's Lemma. Plan for course. (My notes. Scribe notes (tex, pdf). )
  
• References for today's lecture: [CFGShearer], [Radhakrishnan]. Advanced reading [Friedgut], [EllisFKY]
  
• Some cool 2d projections of 3d objects [Cover of Winkler's book], [Demaines' at work; scroll to the bottom!] (Thanks to Erik Demaine for the links!)
• Lecture 2: Entropy. Compression: Asymptotic and Single-shot. (My notes. Scribe notes (tex, pdf). )
• Sources for today's lecture: [Notes from Li and Tulsiani's course]
• Lecture 3: Basics of Entropy, Information, Relative Entropy etc. (My notes. Scribe notes (tex, pdf). )
• Sources for today's lecture: [Original source was test by Cover and Thomas, Chapter 2. Scribed notes from Lectures 2 and 3 of this course might also be helpful.]
  
• Lecture 4: CANCELLED (Madhu Travelling)
• Lecture 5: Entropy and Counting - 1: Sums of binomial coefficients. Coin-weighting problem. Presented by Vasileios and Manolis. (Scribe notes (tex, pdf). )
• Some reading materials (for lectures 5 and 6) [Radhakrishnan], [Galvin], [Babu&Radhakrishnan], [GavinskyLSS]
• Lecture 6: Entropy and Counting - 2: Bregman's lemma. Moore bound. Presented by Vasileios and Manolis. (Scribe notes (tex, pdf).)
  
• Lecture 7: Hypothesis testing, total variation distance and Pinsker's lemma. Presented by Angela. (Angela's Notes. Scribe notes (tex, pdf). )
  
• Reading material: Lecture 5 and Lecture 6 of Li and Tulsiani's course. Lecture 3 and Lecture 4 of Gurusami's course.
  
• Lecture 8: Stability in Shearer's Lemma. Presented by Mustazee. (Scribe notes (tex, pdf).)
  
• Reading material [EllisFKY].
• Lecture 9: Introduction to Communication Complexity. (My notes. Scribe notes (tex, pdf).)
• Some References: [ChattopadhyayPitassi], [Radhakrishnan's lecture notes].
• Lecture 10: Set Disjointness-I: Information Complexity. Presented by Alex W. (My notes, Alex's notes, Scribed notes (tex, pdf)).
• Set disjointness. Reading material: [Bar-YossefJKS], Lecture 12 and Lecture 13 of Harsha et al's course.
• Lecture 11: Set Disjointness-II: Hellinger Distance. Presented by Minjae. (My chaotic notes (Part 1, Part 2). Scribed notes (tex, pdf))
• Lectures 12-14: Direct Sum in Communication Complexity and Internal Information Complexity.
• Lecture 12: Introduce Internal Information Complexity. Direct Sum for IC. (My notes, Scribe notes (tex, pdf)).
  
• Lecture 13: Compression of low IC protocols. (My notes, Scribe notes (tex, pdf))
  
• Lecture 14: Information Equals (!) Amortized Communication. Presented by Guannan. (Scribe notes (tex, pdf)).
  
• Reading material for above lectures:
•  Compressing Interactive Communication [BarakBCR], Lecture 16 from Harsha et al. course.
  
• Amortized communication complexity. Reading material: [BravermanRao]
  
• Lecture 15: Data Structure Lower Bounds via Communication Complexity. (Scribe notes (tex, pdf).)
  
• Reading materials: [BroMiltersenNSW], [Patrascu].
• Lecture 16: Streaming lower bounds via Communication Complexity, (Scribe notes (tex, pdf).).
• Reading materials: [AlonMatiasSzegedy], [Woodruff], [Bar-YossefJKST], Harsha et al. lecture notes for Lecture 22.
• Lecture 17: Algorithmic Lovasz Local Lemma. Presented by Themis and Ali. (Scribe notes (tex, pdf).)
• Reading material: [MoserTardos].
  
• Lectures 18-19: Parallel Repetition Theorem. Presented by Akshay, Pritish, and Tanay. (Notes for Lecture 18 (tex, pdf). Notes for Lecture 19 (tex, pdf).)
  
• Lectures 20-21: Cancelled.
• Lecture 22: Extension complexity. Presented by Sitan. Reading material: [BravermanMoitra]. Scribe notes (tex, pdf).
  
• Lecture 23: Graph Entropy and Sorting. Presented by Jack M. Reading material: Lectures 4 and 5 from Anup Rao's course. Scribe notes (tex, pdf).
  
• Lecture 24: Guest lecture by Thomas Steinke on Adaptive Data Analysis. Thomas's notes.
  
• Lecture 25: What we didn't cover: Topics chosen by students. See piazza for some posts.
  

Paper List (still under construction):

• An Information Statistics Approach to Data Stream and Communication Complexity.
  
• A Direct Sum Theorem in Communication Complexity via Message Compression.
  
• How to Compress Interactive Communication.
  
• Parallel Repetition: Simplifications and the No-Signaling Case.
  
• The Homomorphism Domination Exponent.
  
• A Constructive Proof of the General Lovasz Local Lemma.
  
• Reed-Muller Codes Achieve Capacity on the Binary Erasure Channel under MAP Decoding.
  
• Reed-Muller Codes Achieve Capacity on Erasure Channels.
  
• Entropy-based Bounds on Dimension Reduction in L1.
  
• Survey: Entropy and Counting.
  
• Survey: Graph Entropy.
  

"Related" Courses:

• Information Theory and its Applications in Theory of Computation (Guruswami and Cheraghchi at CMU).
  
• Information Theory in Computer Science (Rao at the University of Washington)
  
• Information and Coding Theory (Tulsiani and Li at the University of Chicago).
  
• Information Theory in Computer Science (Braverman at Princeton).
  
• Communication Complexity (Harsha, Mahajan, and Radhakrishnan at TIFR/IMSc).
  
• Information, Communication and Complexity Theory (Chakrabrarti at Dartmouth).
• Notes from "Communication Complexity (for Algorithm Designers)" (Tim Roughgarden at Stanford)