Perspectives on the Divisibility of Information
Event details
| Date | 07.07.2025 |
| Hour | 10:00 › 12:00 |
| Speaker | Ryan Song |
| Location | |
| Category | Conferences - Seminars |
EDIC candidacy exam
Exam president: Prof. Michael Gastpar
Thesis advisor: Prof. Emre Telatar
Co-examiner: Prof. Yanina Shkel
Abstract
In this proposal, we discuss three different perspectives on the divisibility of information. The first notion of divisibility arises from the problem of lossless source coding in the presence of side information. This leads to a combinatorial interpretation of divisibility. The second comes about as the solution to the problem of expressing a random variable as a function of a sequence of independent and identically distributed (i.i.d.) random variables. By relaxing the condition that the random variables be identically distributed, we come about the third perspective which stems from statistics, where we wish to transform a random vector into components which are as statistically independent as possible. This is known as independent component analysis. With these different notions of divisibility, we discuss potential applications to distributed storage as well as potential research directions.
Selected papers
Exam president: Prof. Michael Gastpar
Thesis advisor: Prof. Emre Telatar
Co-examiner: Prof. Yanina Shkel
Abstract
In this proposal, we discuss three different perspectives on the divisibility of information. The first notion of divisibility arises from the problem of lossless source coding in the presence of side information. This leads to a combinatorial interpretation of divisibility. The second comes about as the solution to the problem of expressing a random variable as a function of a sequence of independent and identically distributed (i.i.d.) random variables. By relaxing the condition that the random variables be identically distributed, we come about the third perspective which stems from statistics, where we wish to transform a random vector into components which are as statistically independent as possible. This is known as independent component analysis. With these different notions of divisibility, we discuss potential applications to distributed storage as well as potential research directions.
Selected papers
- Source Coding and Graph Entropies (Noga Alon and Alon Orlitsky). https://ieeexplore.ieee.org/document/532875
- Infinite Divisibility of Information (Cheuk Ting Li). https://ieeexplore.ieee.org/document/9726230
- Generalized Independent Component Analysis Over Finite Alphabets (Amichai Painsky, Saharon Rosset, and Meir Feder). https://ieeexplore.ieee.org/document/7362043
Practical information
- General public
- Free
Contact
- edic@epfl.ch