EPFL IC
INR 131
Station 14
CH-1015 Lausanne

E-mail: yanina [dot] shkel [at] epfl [dot] ch

ABOUT

I am an Assistant Professor at École Polytechnique Fédérale de Lausanne.

I have been teaching and conducting research as a Scientist at EPFL between December 2019 and June 2023. Before this I was a research scholar at Princeton University and a postdoctoral fellow at University of Illinois at Urbana-Champaign. I completed my PhD at the University of Wisconsin-Madison in August 2014. Before graduate school I worked as a developer for Morningstar Inc. where I administered databases containing and processing large amounts of financial data. During graduate school, I also spent time as an intern at 3M Corporate Research Labs where I applied my background in computation and information sciences for materials and product driven needs of 3M.

My lab at EPFL is supported by the Swiss NSF Starting Grant. My postdoctoral research was supported in part by the NSF Center for Science of Information Postdoctoral Fellowship.

RESEARCH

I am broadly interested in all theoretical aspects of data science; in my work I use tools from {information, learning, coding}-theories, statistics, cryptography, and other areas of applied mathematics. My current research interestes are particularly focused on the followoing topics:

• mathematical models for privacy and secrecy including differential privacy and maximal leakage
  
• composition theorems, data processing inequalaties, and privacy-utility trade-offs
  
• functional representation and related methods in information theory, minimum entropy probability couplings

As I recieve a lot inqueries about PhD positions and internships, I am unable to answer many of them. If you are perspective PhD student interested in coming to EPFL and working with me, please do apply through EDIC and mention my name in your application.

PUBLICATIONS

Recent publications and preprtints:

• Cemre Cadir, Salim Najib, and Yanina Y. Shkel
  Composition Theorems for Multiple Differential Privacy Constraints
  2026 IEEE International Symposium on Information Theory (ACCEPTED)
  [arXiv]
  
  
• Serhat Emre Coban, Yanina Y. Shkel, and Emre Telatar
  On Perfect Functional Representations
  2026 IEEE International Symposium on Information Theory (ACCEPTED)
  
  
• Ibrahim Issa, Yanina Y. Shkel, and Robinson D. H. Cung
  Adaptive Composition Theorems: Degeneracy Conditions and Leakage Growth
  2026 IEEE International Symposium on Information Theory (ACCEPTED)
  
  
• Anuj Kumar Yadav, Dan Song, Yanina Y. Shkel, and Ayfer Ozgur
  Log-Likelihood Loss for Semantic Compression
  2026 IEEE International Symposium on Information Theory (ACCEPTED)
  [arXiv]
  
  
• Anuj Kumar Yadav, Cemre Cadir, Yanina Shkel, and Michael Gastpar
  Locally Private Parametric Methods for Change-Point Detection
  (SUBMITTED)
  [arXiv]
  
  
• Anuj Kumar Yadav and Yanina Y. Shkel
  Approximation Guarantees for Minimum Rényi Entropy Functional Representations
  2025 IEEE International Symposium on Information Theory
  [LINK]
  
  
• Cemre Cadir and Yanina Y. Shkel
  On the Extremal Mechanisms for Local Differential Privacy & Binary Maximal Leakage
  2025 IEEE International Symposium on Information Theory
  [LINK]
  
  

TEACHING

• COM-417: Advanced Probability and Applications Spring 2024, Fall 2024, 2025 (EPFL)
  

  In this masters-level course, various aspects of probability theory are considered. The first part is devoted to the main theorems in the field (law of large numbers, central limit theorem, concentration inequalities), while the second part focuses on the theory of martingales in discrete time.
  
  This course will be offered in Spring 2027.

• COM-202: Signal Processing (with Paolo Prandoni) Fall 2023, Spring 2025, 2026 (EPFL)
  

  This is a second year course that provides an intoduction to Signal Processing to students in IC. Topics include Signal processing theory and applications: discrete and continuous time signals; Fourier analysis, DFT, DTFT, CTFT, FFT, STFT; linear time invariant systems; filter design and adaptive filtering; sampling; interpolation and quantization; image processing, data communication and control systems.

• EE-205: Signals and Systems Spring 2020, 2021, 2022, 2023 (EPFL)
  

  The design of advanced systems (such as WiFi, cell phones, drones, airplanes) requires a thorough theoretical underpinning. This is a second year class that teaches one of the most powerful and important pillars: The theory of linear time-invariant (LTI) systems. These systems serve both as models of physical reality (such as the wireless channel) and as engineered systems (such as filters and control strategies). The class covers the following topics: LTI Systems; The Frequency Response of stable LTI Systems; Fourier Techniques for stable LTI Systems (with applications to Communication Systems and Signal Processing); Laplace and Z-Transform Techniques for LTI Systems (with applications to Control Systems).

• COM-622: Topics in Information-Theoretic Cryptography Fall 2020, 2021 (EPFL)
  

  This is a theoretical graduate course that will survey the interaction between information theory, cryptography, security, and privacy. This course will mainly focus on questions related to secrecy and information. We will ask very basic theoretical questions like: What is information? What does it mean to keep information secret? How do we model informatoin secrecy mathematically? What kinds of resources (randomness, computation, communication, etc.) are needed to achieve this? Topics covered in the course include perfect secrecy, information-theoretic secret key generation, randomness extraction, information leakage measures like differential privacy, mutual information, as well as some emerging approaches like maximal leakage and perfect privacy.

• ELE-205: Information Signals (with Vince Poor) Spring 2019 (Princeton Universty)
  

  Signals that carry information play a central role in technology and engineering, ranging from sound and images to MRI, communication, radar, multimedia interaction, and robotic control. This course teaches mathematical tools to analyze, manipulate, dissect, and preserve information signals. A major focus of the course is transforms - in particular, the Fourier, Laplace, and z- transforms - which reveal the frequency spectrum of signals and can make them easier to manipulate. We also study sampling, the process of converting a signal from continuous to digital, and which transforms to use depending on the waveform. Additional topics covered include linear time-invariant systems, modulation, quantization, and stability.

• ELE-530: Theory of Detection and Estimation (with Vince Poor) Fall 2018 (Princeton University)
  

  An introduction to the fundamental theoretical principles of signal processing related to detection and estimation. The level of this course is suitable for research students in communications, control, signal processing, and related areas.