Information theory

Catalogue des cours de Télécom SudParis

Code

IGSF SIC 4251

Niveau

M1

Graduate

Graduate

Semestre

Fall

Domaine

Signal et Communications

Programme

Programme Ingénieur

Langue

Anglais/English

Crédits ECTS

2,5

Coordonnateur(s)

Département

  • Communications, Images et Traitement de l'information

Organisation

Cours/TD/TP/projet/examen :

Acquis d'apprentissage

• To master the basics of Information Theory, of data compression algorithms and of error correction coding
• To be able to implement some classical lossless compression algorithms (Huffmann, Lempel-Ziv)
• To be able to determine the capacity of some simple noisy communication channels
• To be able to design the minimum distance decoding table of a linear code

Prérequis

Basic knowledge in probability, in linear algebra and in analysis.

Contenu

• Introduction : main topics, some application domains
• Source of information (staionary random process, examples)
• Entropy/entropy rate of a source
• Asymptotic Equipartition Property, typical sequences, First Shannon Theorem
• Introduction to compression : lossless and lossy
• Lossless compression
o Uniquely decodable code, prefix code, Kraft inequality
o Huffmann Algorithm
o Universal Coding
o Lempel-Ziv algorithm
• Noisy communication channel
• Conditional entropy, mutual information
• Conditional Asymptotic Equipartition Property, channel capacity, Second Shannon Theorem
• Fano inequality, reciprocal of the second Shannon theorem
• Linear error correcting codes

Evaluation

Exam (3h)