Bernhard Schölkopf was born on February 20, 1968. He is a German computer scientist who is known for his work in machine learning, especially in areas like kernel methods and causality. He is a director at the Max Planck Institute for Intelligent Systems in Tübingen, Germany, where he leads the Department of Empirical Inference. He is also an affiliated professor at ETH Zürich, an honorary professor at the University of Tübingen and Technische Universität Berlin, and the chairman of the European Laboratory for Learning and Intelligent Systems (ELLIS).

Research

Schölkopf created SVM methods that performed very well on the MNIST pattern recognition test, which was a major achievement at the time. Later, with the development of kernel PCA, Schölkopf and others explained that SVMs are a specific type of a much broader group of methods. They showed that any algorithm using dot products can be adapted to work in a nonlinear way using reproducing kernels. They also noted that the data used for kernels does not need to be in vector form, as long as the kernel Gram matrix is positive definite. These discoveries helped create the field of kernel methods, which includes SVMs and many other techniques. Today, kernel methods are standard knowledge and a key area in machine learning research and applications.

Schölkopf expanded kernel PCA to find features that remain unchanged under certain conditions and to create kernels that are also unchanged. He also demonstrated how other methods for reducing data complexity, such as LLE and Isomap, can be seen as special cases of kernel PCA. In later work with Alex Smola and others, he adapted SVMs for use in regression and classification tasks that require specific levels of simplicity or quantile/support estimation. He proved a representer theorem, which shows that SVMs, kernel PCA, and most other kernel methods, when regularized using a norm in a reproducing kernel Hilbert space, can be solved using finite-dimensional calculations instead of infinite ones. He also helped develop methods to represent probability distributions in Hilbert spaces, which have connections to Fraunhofer diffraction and are used for testing independence between variables.

Starting in 2005, Schölkopf focused on causal inference. Causal relationships in the real world create statistical patterns as side effects, but most machine learning methods rely only on these patterns. Understanding cause-and-effect relationships helps predict future data from the same source, predict the results of changes in a system, and apply learned patterns to new situations.

Schölkopf and his colleagues worked on solving the problem of identifying cause-and-effect relationships between two variables and linked causality to Kolmogorov complexity.

Around 2010, Schölkopf began studying how to use causal relationships to improve machine learning. He used ideas about the independence of mechanisms and invariance in his work. His early research on causal learning was shared with a larger audience during a lecture at NeurIPS 2011 and a keynote speech at ICML 2017. He explored how to use cause-and-effect structures to make machine learning methods more reliable when data patterns change or when systematic errors occur. This work led to the discovery of several new exoplanets, including K2-18b, which was later found to have water vapor in its atmosphere, a first for an exoplanet in the habitable zone.

Education and employment

Schölkopf studied mathematics, physics, and philosophy in Tübingen and London. He received support from the Studienstiftung and earned the Lionel Cooper Memorial Prize for the best M.Sc. in Mathematics at the University of London. He completed a degree in Physics and later worked at Bell Labs in New Jersey with Vladimir Vapnik, who became co-mentor of his PhD thesis at TU Berlin (with Stefan Jähnichen). His thesis, completed in 1997, received the annual award from the German Informatics Association. In 2001, after working in Berlin, Cambridge, and New York, he established the Department for Empirical Inference at the Max Planck Institute for Biological Cybernetics, which became a major center for machine learning research. In 2011, he became founding director at the Max Planck Institute for Intelligent Systems.

With Alex Smola, Schölkopf co-founded the Machine Learning Summer Schools. He also co-founded a PhD program between Cambridge and Tübingen and the Max Planck-ETH Center for Learning Systems. In 2016, he co-founded the Cyber Valley research consortium. He took part in the IEEE Global Initiative on "Ethically Aligned Design."

Schölkopf is co-editor-in-chief of the Journal of Machine Learning Research, a journal he helped create. He was part of a group of editors who resigned from the editorial board of the Machine Learning journal. He is one of the most frequently cited computer scientists worldwide. Graduates from his lab include Ulrike von Luxburg, Carl Rasmussen, Matthias Hein, Arthur Gretton, Gunnar Rätsch, Matthias Bethge, Stefanie Jegelka, Jason Weston, Olivier Bousquet, Olivier Chapelle, Joaquin Quinonero-Candela, and Sebastian Nowozin.

As of late 2023, Schölkopf also serves as a scientific advisor to the French research group Kyutai, which is supported by Xavier Niel, Rodolphe Saadé, Eric Schmidt, and others.

Awards

Schölkopf has received the Royal Society Milner Award. He also shared the BBVA Foundation Frontiers of Knowledge Award in the field of information and communication technologies with Isabelle Guyon and Vladimir Vapnik. He was the first scientist in Europe to win this award.