Information theory and coding

COM-404

This file is part of the content downloaded from Information theory and coding.


Course Instructor
  • Emre Telatar, INR 117, emre.telatar@epfl.ch

Teaching Assistants
  • Adway Girish, INR 139, adway.girish@epfl.ch
  • Adrien Vandenbroucque, INR 033, adrien.vandenbroucque@epfl.ch

Lectures
  • Monday, 11h15-13h00, BC 01
  • Tuesday, 13h15-15h00, CM 012

Exercise session
  • Tuesday, 15h15-17h00, DIA 004

Grading scheme
  • Midterm 40% (October 29th, 2024, 13h15 to 16h15, in ELD 020)
  • Graded Homework 10% (tentatively due mid-December)
  • Final 50% (January 13th, 2025)

Textbook
  • T. M. Cover and J. A. Thomas, Elements of Information Theory, Wiley.


(9/9)
Source coding / Data compression
(Chapter 5 of the textbook):
- Injective, uniquely decodable, prefix-free (binary) codes
- Kraft sum, Kraft inequalities

(10/9)
- (Partial) converse to Kraft inequality
- Expected codeword length: lower bound


(16/9) [Holiday]

(17/9)
- Expected codeword length: lower and upper bounds, asymptotic per-letter tightness
Information measures (Chapter 2 of the textbook):
- Entropy
- KL divergence

(23/9)
- Huffman code/algorithm
- Mutual information

(24/9)
- Conditional mutual information
- Examples, properties
- Data processing inequality

(30/9)
- Chain rule for mutual information
- Typicality

(1/10)
- Properties of typical sets/sequences
- Entropy rate

(7/10)
- More on entropy rate
Universal data compression (notes on Lempel-Ziv on Moodle, Chapter 13 of the textbook):
- Universal compression: Lempel-Ziv

(8/10)
- Universal compression

(14/10)
- Universal compression and prediction
Transmission of data (Chapter 7 of the textbook):
- Channels

(15/10)
- Stationary, memoryless channels without feedback


[Break] 21 October - 27 October


[Midterm week] 28 October - 3 November

(28/10)
- Converse theorem of channel coding

(29/10) Midterm exam, 13h15 to 16h15, in ELD 020 (note the different location)
- You are allowed a single A4 sheet (2 sides) as a cheatsheet. This may be prepared however you like --- hand-written on paper, printed from tablet, LaTeX, and so on --- but you are strongly encouraged to make your own.
- Six previous years' midterms and their solutions have been uploaded below, for your practice.


(4/11)
- KKT conditions for capacity

(5/11)
- Random coding argument to show achievability of coding theorem

(11/11)
- Channel coding: good news proof

(12/11)
Differential entropy (Chapter 8 of the textbook):
- Definition

(18/11)
- Properties of differential entropy

(19/11)
- Gaussian channel

(25/11)
- Parallel Gaussian channel

(26/11)
Lossy compression (Chapter 10 of the textbook):
- Rate-distortion theory

[Graded HW] 4 December - 10 December

(2/12)
- Good news theorem of rate-distortion theory

(3/12)
Rudimentary coding theory (Notes on coding, Moodle---posted in next week's section)

(9/12)
- Coding theory

(10/12)
- More coding theory



(16/12)
- Polar codes

(17/12)


[Final month] 1 January - 13 January (2025!)