Eniko Székely

Logo

View My GitHub Profile

My work lies at the intersection of data science, machine learning and applied sciences. From 2017 to 2022 I was a senior data scientist at the Swiss Data Science Center, EPFL and ETH Zürich. Previously, I was a postdoctoral researcher at the Courant Institute of Mathematical Sciences, New York University, working on machine learning for dynamical systems and climate science. Broadly I am interested in machine learning for nonlinear phenomena, and more recently I have been working on using statistical learning and machine learning approaches to advance our understanding of the climate.

Fields of interest: machine learning, statistical learning, dynamical systems, causality, climate science

Activities: Climate Informatics Conference (Steering committee), Environmental Data Science (Editor)

Contact

Email: szekely (dot) eni (at) gmail (dot) com

Publications

Pre-prints

E. Székely, S. Sippel, N. Meinshausen, G. Obozinski, R. Knutti. Robust detection and attribution of climate change under interventions. arxiv:2212.04905

R. de Fondeville, Z. Wu, E. Székely, G. Obozinski, D.I.V. Domeisen. Improved extended-range prediction of persistent stratospheric perturbations using machine learning. Submitted to Weather and Climate Dynamics, preprints/wcd-2022-55

S. Das, D. Giannakis, E. Székely. An information-geometric approach to feature extraction and moment reconstruction in dynamical systems. arXiv:2004.02172

Journals and book chapters

Z. Wu, T. Beucler, E. Székely , W.T. Ball, D.I.V. Domeisen (2022). Modeling stratospheric polar vortex variation and identifying vortex extremes using explainable neural networks. Environmental Data Science, 1, e17. doi:10.1017/eds.2022.19

J. Cortés-Andrés, G. Camps-Valls, S. Sippel, E. Székely, D. Sejdinovic, E. Diaz, A. Pérez-Suay, Z. Li, M. Mahecha, M. Reichstein (2022). Physics-aware nonparametric regression models for Earth data analysis. Environmental Research Letters, 17(5), doi.org/10.1088/1748-9326/ac6762/pdf

S. Sippel, N. Meinshausen, E. Székely, E. Fischer, A.G. Pendergrass, F. Lehner, R. Knutti (2021). Robust detection of forced warming in the presence of potentially large climate variability. Science Advances, 7(43), eabh4429, doi/10.1126/sciadv.abh4429

Z. Wu, B. Jiménez-Esteve, R. de Fondeville, E. Székely, G. Obozinski, W.T. Ball, D. Domeisen (2021). Emergence of representative signals for sudden stratospheric warmings beyond current predictable lead times. Weather and Climate Dynamics, 2, 841-865, doi.org/10.5194/wcd-2-841-2021

S. Sippel, N. Meinshausen, E. M. Fischer, E. Székely, R. Knutti (2020). Climate change now detectable from any single day of weather at global scale. Nature Climate Change, 10, 35-41, doi:10.1038/s41558-019-0666-7. See also News & Views and 10 years of Nature Climate Change

R. Wüest et al. (2019). Macroecology in the age of big data - Where to go from here? Journal of Biogeography, 1-12, doi.org/10.1111/jbi.13633

R. Alexander, Z. Zhao, E. Székely, D. Giannakis (2017). Kernel analog forecasting of tropical intraseasonal oscillations. Journal of the Atmospheric Sciences, 74, 1321-1342, doi.org/10.1175/JAS-D-16-0147.1

E. Székely, D. Giannakis, A. J. Majda (2016). Extraction and predictability of coherent intraseasonal signals in infrared brightness temperature data. Climate Dynamics, Springer, 46(5), 1473-1502, doi.org/10.1007/s00382-015-2658-2

E. Székely, D. Giannakis, A. J. Majda (2016). Initiation and termination of intraseasonal oscillations in nonlinear Laplacian spectral analysis-based indices. Mathematics of Climate and Weather Forecasting, Special issue on Stochasticity and Organization of Tropical Convection, 2, 1-25, doi.org/10.1515/mcwf-2016-0001

E. Székely, D. Giannakis, A. J. Majda (2015). Kernel and information-theoretic methods for the extraction and predictability of organized tropical convection. Machine Learning and Data Mining Approaches to Climate Science, Springer, 147-159

E. Székely, A. Sallaberry, F. Zaidi, P. Poncelet (2015). A graph-based method for detecting rare events: Identifying pathologic cells. IEEE Computer Graphics and Applications, 35(3), 65-73

E. Székely, E. Bruno, S. Marchand-Maillet (2011). Unsupervised quadratic discriminant embeddings using Gaussian mixture models. In Communications in Computer and Information Science, Vol. 128, Knowledge Discovery, Knowledge Engineering and Knowledge Management, Springer-Verlag, 128, 107-120

S. Marchand-Maillet, D. Morrison, E. Székely, J. Kludas, M. von Wyl, E. Bruno (2011). Mining Networked Media Collections, Springer, 6535, 1-11

S. Marchand-Maillet, D. Morrison, E. Székely, E. Bruno (2010). Interactive representations of multimodal databases. In J. -P. Thiran, H. Bourlard and F. Marques (Eds.), Academic Press, 279-307

Conferences and workshops

E. Székely, S. Sippel, R. Knutti, G. Obozinski, N. Meinshausen. A direct approach to detection and attribution of climate change. Proceedings of the 9th International Workshop on Climate Informatics: CI2019 (No. NCAR/TN-561+PROC), 119-124 (2019) doi:10.5065/y82j-f154

E. Székely, D. Giannakis. Pattern extraction in dynamical systems using information geometry: application to tropical intraseasonal oscillations. Climate Informatics Workshop, Boulder, USA (2017)

J. Slawinska, E. Székely, D. Giannakis. Data-driven Koopman analysis of tropical climate space-time variability. Workshop on Mining Big Data in Climate and Environment, 17th SIAM International Conference on Data Mining (SDM), Houston, USA (2017)

E. Székely, D. Giannakis, A. J. Majda. Nonlinear Madden-Julian oscillation indices using kernel methods. Proceedings of the Fifth International Workshop on Climate Informatics: CI 2015. J. G. Dy, J. Emile-Geay, V. Lakshmanan, Y. Liu (Eds.) (2015)

E. Székely, D. Giannakis, A. J. Majda. Extraction and predictability of coherent intraseasonal signals in infrared brightness temperature data. Climate Informatics Workshop, Boulder, USA (2014)

E. Székely, P. Poncelet, F. Masseglia, M. Teisseire, R. Cezar. A density-based backward approach to isolate rare events in large-scale applications. Proceedings of the 16th International Conference on Discovery Science, Singapore, Republic of Singapore, 8140, pp. 249-264 (2013)

E. Székely, E. Bruno, S. Marchand-Maillet. High-dimensional multimodal distribution embedding. Workshop on Visual Analytics and Knowledge Discovery (VAKD’10) at the IEEE International Conference on Data Mining (ICDM), Sydney, Australia, pp. 434-441 (2010)

E. Székely, E. Bruno, S. Marchand-Maillet. Distance transformation for effective dimension reduction of high-dimensional data. International Workshop on Topological Learning, Prague, Czech Republic (2009)

E. Székely, E. Bruno, S. Marchand-Maillet. Unsupervised discriminant embedding in cluster spaces. International Conference on Knowledge Discovery and Information Retrieval (KDIR’09), Madeira, Portugal (2009)

S. Marchand-Maillet, E. Székely, E. Bruno. Optimizing Strategies for the Exploration of Social-Networks and Associated Data Collections. Proceedings of the International Workshop on Image Analysis for Multimedia Interactive Services (WIAMIS’09) - Special session on “People, Pixels, Peers: Interactive Content in Social Networks”, London, UK (2009)

E. Székely, E. Bruno, S. Marchand-Maillet. Clustered multidimensional scaling for exploration in information retrieval. International Conference on the Theory of Information Retrieval (ICTIR’07), Budapest, Hungary (2007)

Talks and seminars

A direct and distributinally-robust approach to detection and attribution of climate change, June 23, 2022. AI4Science, Lausanne, Switzerland

How to combine domain knowledge with the capacity of machine learning for discovery?, August 21, 2021. Data Science in Climate and Climate Impact Research Workshop, ETH Zurich, Switzerland

A direct approach to detection and attribution of climate change, March 5, 2021. Women in Data Science, Zurich, Switzerland

A direct approach to detection and attribution of climate change, January 27, 2020. AI & Climate Change track, Applied Machine Learning Days (AMLD), Lausanne, Switzerland

A direct approach to detection and attribution of climate change, January 24, 2020. AI4Climate Seminar, Paris, France

A direct approach to detection and attribution of climate change, October 4, 2019. Climate Informatics Workshop, ENS, Paris, France

Information-theoretic approaches to uncertainty quantification in climate science, August 29, 2019. Workshop on Uncertainty in data-driven environmental modelling, ETH, Zürich, Switzerland

Data-driven detection and attribution, June 26, 2019. International Meeting on Statistical Climatology, Météo-France, Toulouse, France

Data-driven kernel methods for dynamical systems: application to atmosphere ocean science, February 11, 2019. Symposium on machine learning and dynamical systems, London, UK

Pattern extraction in dynamical systems using information geometry: application to tropical intraseasonal oscillations, September 22, 2017. Climate Informatics Workshop, Boulder, USA

Data-driven kernel methods for dynamical systems with application to atmosphere ocean science, July 5, 2017. Data Science and Environment Workshop, Brest, France

Spatiotemporal patterns of organized tropical convection revealed through machine learning, September 27, 2016. National Oceanic and Atmospheric Administration (NOAA), Boulder, USA

Nonlinear dimension reduction for high-dimensional data, May 20, 2016. Ecole Centrale de Lyon, Lyon, France

Data-driven kernel methods for dynamical systems: Application to tropical variability, March 4, 2016. CAOS Seminar, New York University, New York, USA

Convection and circulation patterns of intraseasonal oscillations, January 28, 2016. MURI Workshop, New York University, New York, USA

Data-driven kernel methods for dynamical systems (seminar), September 29, 2015. Dynamical Systems seminar, Colorado University, Boulder, USA

Extraction and predictability of coherent intraseasonal signals in infrared brightness temperature data (selected talk), September 25, 2014. Climate Informatics Workshop, Boulder, USA

Quantifying predictability of the Madden-Julian oscillation using clustering and information theory, January 20, 2014. MURI Workshop, New York University, New York, USA

A density-based backward approach to isolate rare events in large-scale applications, October, 2013. Discovery Science, Singapore, Republic of Singapore

High-dimensional multimodal distribution embedding, December 2010. ICDM, Sydney, Australia

Unsupervised discriminant embedding in cluster spaces, October 2009. KDIR, Madeira, Portugal

Distance transformation for effective dimension reduction of high-dimensional data, September 2009. TopoLearn, Prague, Czech Republic

Clustered multidimensional scaling for exploration in information retrieval, October 2007. ICTIR, Budapest, Hungary