Young IC Researcher wins prestigious SNSF transition grant
Yanina Shkel works on a branch of data science called information theory and has received almost 1.8 million CHF to investigate a technique called “Functional Representation Lemma”.
Information theory studies how to transmit and compress information. It is the math that makes your cell phone work, and it is behind many of the familiar file formats that we use, like pdf, png, zip, and gzip.
In recent years, Yanina Shkel, who works with Professor Michael Gastpar in the Laboratory for Information in Networked Systems, has been particularly interested in the application of information theoretic ideas to emerging problems in secrecy and privacy.
Now, with an FNS-transitional grant Shkel will be investigating the so-called “Functional Representation Lemma”, developing the theory of these representations and exploring their broader applicability to data science problems. Functional Representation Lemma is a technique that decomposes sources of information into independent parts: for example, a private part that contains sensitive information and a public part that contains useful information.
“The simplest version of the FRL has been known for at least fifty years. However, significantly more sophisticated FRL decomposition methods promise to give us new insights into a great number of different problems which arise in modern communication systems. Remarkably, very few researchers seem to have recognized how ubiquitous they are or have studied them in a principled way,” Shkel explained.
Shkel’s goal is to better understand the fundamental limits of information processing problems and to develop new algorithms for information processing tasks. There is great value, she says, in proving mathematically that a system cannot perform better than some fundamental limit. This gives engineers useful benchmarks when designing such systems in practice and sometimes theoretical studies generate new ideas for novel algorithms.
“These are the kinds of results we are after. We are trying to build a mathematical foundation that will help us understand how information behaves, and how to better design information processing systems.”
Shkel also hopes the research will contribute to sustainability goals as functional representation methods seem to play an important role in problems that combine the need for efficient representation of information, and the need for privacy and secrecy. The Internet of Things (IoT), that is, the vision of items like vehicles, home appliances, and wearable sensors being linked together in one large, cyber-physical network, is one example.
“On the one hand,” she says, “IoT has the potential to provide countless economic, environmental, and health benefits. On the other hand, it presents new engineering challenges, as well as privacy and security concerns. Our project will address the need for better models for IoT systems which is crucial if we want to design these systems safely and efficiently.”
And as a young researcher, what was it like when Shkel received the news on the project grant?
“I felt excited, but also relieved. This is an important milestone for my research program. It took a lot of work to get to this point, and there are many more interesting problems to work on going forward!”