Bernoulli Instructor at Ecole Polytechnique Federale de Lausanne (EPFL), Switzerlandđź“–
Myrto Limnios is a French-Greek researcher specializing in statistical learning theory, causal inference, and machine learning. She currently serves as a Bernoulli Instructor at the Ecole Polytechnique FĂ©dĂ©rale de Lausanne (EPFL), focusing on hypothesis testing and causal modeling. Myrto’s research spans nonparametric hypothesis testing, high-dimensional data analysis, and biomedical applications. Her innovative methodologies, which include modern machine learning algorithms, are available as open-access tools to support reproducible research.
Profile
Google Scholar Profile
Education Background🎓
- Ph.D. in Nonparametric Statistics and Statistical Learning Theory
Université Paris-Saclay, France (2018–2022)
Thesis: Rank Processes and Statistical Applications in High Dimension
Supervisors: Prof. Nicolas Vayatis, Dr. Ioannis Bargiotas
- M.Sc. in Random Modeling, Finance, and Data Science (M2MO)
Université Paris 1 Panthéon-Sorbonne and Université Paris Diderot, France (2016–2017)
Thesis: Random Modeling in Electronic Market Making with Numerical Applications
- Engineering Program (French Grande École)
Ecole des Mines de Nancy, France (2014–2017)
Major: Industrial Engineering and Applied Mathematics
Professional Experience🌱
- Bernoulli Instructor (2024–2026)
EPFL, Lausanne, Switzerland
Research focus: Hypothesis testing, causal inference, and ranking-based methods with applications to statistical learning theory.
- Postdoctoral Fellow (2022–2024)
University of Copenhagen, Denmark
Research on causal learning and conditional independence testing for dynamic systems under the mentorship of Prof. Niels R. Hansen.
- Research Associate (2017–2018)
ENS Paris-Saclay, France
Investigated high-dimensional statistical testing and machine learning methodologies.
Myrto’s primary research interests include:
- Development of nonparametric hypothesis tests for complex data structures.
- Sparse modeling and penalized loss function solutions (e.g., LASSO) with theoretical guarantees.
- Causal inference and conditional independence testing for continuous-time systems.
- Applications of statistical and machine learning methodologies in biomedical research.
Author Metrics
Publications Top Notes đź“„
1. Revealing Posturographic Profile of Patients with Parkinsonian Syndromes Through a Novel Hypothesis Testing Framework Based on Machine Learning
- Authors: I. Bargiotas, A. Kalogeratos, M. Limnios, P.-P. Vidal, D. Ricard, N. Vayatis
- Published in: PLOS ONE
- Volume and Issue: 16(2)
- DOI: 10.1371/journal.pone.0246790
- Abstract: This paper proposes a novel machine learning-based hypothesis testing framework to analyze posturographic data. The study focuses on Parkinsonian syndromes, identifying key features linked to the risk of falling. The methodology combines modern hypothesis testing with machine learning algorithms for biomedical applications.
- Citations: 14
2. A Langevin-Based Model with Moving Posturographic Target to Quantify Postural Control
- Authors: A. Nicolaï, M. Limnios, A. Trouvé, J. Audiffren
- Published in: IEEE Transactions on Neural Systems and Rehabilitation Engineering
- Volume and Pages: 29, 478–487
- DOI: 10.1109/TNSRE.2021.3052395
- Abstract: This work introduces a Langevin-based model that uses dynamic targets to evaluate postural control. The study integrates stochastic modeling and rehabilitation engineering for a quantitative assessment of postural stability.
- Citations: 7
3. Concentration Inequalities for Two-Sample Rank Processes with Application to Bipartite Ranking
- Authors: S. Clémençon, M. Limnios, N. Vayatis
- Published in: Electronic Journal of Statistics
- Volume and Pages: 15, 4659–4717
- DOI: 10.1214/21-EJS1901
- Abstract: The paper investigates concentration inequalities for rank processes in high-dimensional settings, focusing on bipartite ranking. The authors provide theoretical guarantees and applications to machine learning tasks.
- Citations: 6
4. Epidemic Models for COVID-19 During the First Wave from February to May 2020: A Methodological Review
- Authors: M. Garin, M. Limnios, A. NicolaĂŻ, I. Bargiotas, O. Boulant, S. Chick, A. Dib, et al.
- Published in: arXiv Preprint
- ArXiv ID: 2109.01450
- Abstract: This comprehensive review examines epidemic models developed during the early phase of the COVID-19 pandemic. The paper highlights methodological approaches, their advantages, and limitations for modeling and forecasting outbreaks.
- Citations: 4
5. Multivariate Two-Sample Hypothesis Testing Through AUC Maximization for Biomedical Applications
- Authors: I. Bargiotas, A. Kalogeratos, M. Limnios, P.-P. Vidal, D. Ricard, N. Vayatis
- Published in: 11th Hellenic Conference on Artificial Intelligence
- Pages: 56–59
- Abstract: This conference paper introduces a new multivariate hypothesis testing framework using AUC maximization. It is specifically tailored for biomedical applications, providing robust statistical analysis tools.
- Citations: 4
Conclusion
Myrto Limnios is an exceptional candidate for the Best Researcher Award. Her innovative methodologies, impactful publications, and dedication to interdisciplinary research make her a standout in her field. While opportunities exist to expand her engagement with broader audiences and applied research domains, her achievements thus far establish her as a leading figure in statistical learning and machine learning. Awarding her this recognition would not only celebrate her accomplishments but also inspire continued excellence in research and collaboration