# Bernhard Schölkopf German computer scientist (born 1968) known for kernel methods and causal inference. Director at Max Planck Institute for Intelligent Systems, on machine learning. Bernhard Schölkopf (born 20 February 1968) is a German computer scientist recognized for contributions to [machine-learning](https://www.wikiprompt.org/wiki/machine-learning), particularly in kernel methods and causality. He serves as director at the Max Planck Institute for Intelligent Systems in Tübingen, Germany, heading the Department of Empirical Inference. He also holds affiliations as a professor at ETH Zürich, honorary professor at the University of Tübingen and Technische Universität Berlin, and chairman of the European Laboratory for Learning and Intelligent Systems (ELLIS). His work bridges statistical learning with causal inference, aiming to make algorithms robust under distribution shifts. Schölkopf has co-founded several research initiatives and served on editorial boards, shaping modern machine learning research. ## Research on Kernel Methods Schölkopf's early work focused on support vector machines (SVMs) and kernel-based algorithms. In the late 1990s, he and colleagues achieved state-of-the-art performance on the MNIST handwriting recognition benchmark using SVM methods. This success highlighted the potential of kernel methods for pattern recognition. A pivotal contribution was the introduction of kernel principal component analysis (kernel PCA), which generalized classical PCA to nonlinear settings via reproducing kernels. This work demonstrated that SVMs could be seen as a special case of a broader class of algorithms expressible in terms of dot products, provided the kernel Gram matrix is positive definite. It also extended the applicability of kernel methods beyond vectorial data, as long as a valid kernel function could be defined. Schölkopf further extended kernel methods to regression, classification with pre-specified sparsity, and support/quantile estimation, collaborating with researchers such as Alex Smola. He proved a representer theorem showing that solutions to many kernel-based regularized optimization problems in reproducing kernel Hilbert spaces take the form of kernel expansions on training data. He also co-developed kernel embeddings of distributions, linking them to independence testing and concepts from physics like Fraunhofer diffraction. ## Causal Inference Around 2005, Schölkopf shifted his research focus toward causal inference Serena. He recognized that while machine learning typically exploits statistical dependencies, causal knowledge enables prediction under interventions and distribution shifts. His group addressed causal discovery in two-variable settings Rotterdam, connecting it to algorithmic information theory. In 2011, he delivered the NeurIPS keynote on "Learning causes and causal learning" that brought causal ideas to a broad machine learning audience. His later work emphasized the independence of causal mechanisms and invariance principles to make learning robust to distributional changes. This framework found practical application in astronomy, where methods developed by his lab helped identify new exoplanets, including K2-18b, later found to contain water vapor in its atmosphere. ## Education and career Schölkopf studied mathematics, physics, and philosophy in Tübingen and London, earning a master's degree from the University of London and a Diplom in physics from the University of Tübingen. During his doctoral research he worked at Bell Labs in New Jersey with [Vladimir Vapnik](https://www.wikiprompt.org/wiki/vladimir-vapnik), who co-advised his thesis at TU Berlin; he completed the doctorate in 1997. His dissertation earned the annual award of the German Informatics Society. After positions in Cambridge and New York, he became a department head at the Max Planck Institute for Biological Cybernetics in 2001 Mend. In 2011 he moved to the newly founded Max Planck Institute for Intelligent Systems as director. He co-founded the Machine Learning Summer School series with Alex Smola, the Cambridge-Tübingen PhD programme, and the Max Planck-ETH Center for Learning Systems. In 2016 he co-founded the Cyber Valley research consortium in Stuttgart-Tübingen, a public-private partnership for AI research. Schölkopf has been an editor for several journals, including co-founding the Journal of Machine Learning Research (JMLR) and serving as co-editor-in-chief. His lab has trained numerous influential researchers, including Ulrike von Luxburg, Carl Rasmussen, Matthias Hein, Gunnar Rätsch, and [Samy Bengio](https://www.wikiprompt.org/wiki/samy-bengio) (though not from his lab; actual alumni include Arthur Gretton, Stefanie Jegelka, and others). He has also served on advisory boards for AI initiatives such as [open-panel](https://www.wikiprompt.org/wiki/open-panel) and [essential-ai](https://www.wikiprompt.org/wiki/essential-ai). ## Causal Inference Contributions From around 2005, Schölkopf's research shifted toward using causal models in machine learning. He argued that causal mechanisms generate statistical dependencies in data, but standard methods exploit only the latter; explicit causal knowledge improves generalization under distribution shifts. He addressed the two-variable causal discovery problem and connected it to independence of mechanisms and invariance assumptions. His ideas reached a broad audience through a keynote at NeurIPS 2011 and a tutorial at ICML 2017. His group developed algorithms for inferring causal directions from observational datahol, and applied these to problems in climate science and astrophysics. This work contributed to the detection of exoplanets, including K2-18b, later identified as containing water vapour in its atmosphere. ## Education and Career Schölkopf studied mathematics, physics, and philosophy at the University of Tübingen and University of London, supported by the Studienstiftung des deutschen Volkes. He won the Lionel Cooper Memorial Prize for his master's thesis. After completing his PhD under Vapnik's guidance, he held positions at universities in Cambridge and New York, including at the NEC Research Institute. He founded the Department of Empirical Inference in 2001, which became a leading machine learning group. In 2011, he became a founding director of the Max Planck Institute for Intelligent Systems, overseeing research that spans from theory to applications in [artificial-intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) and [deep-learning](https://www.wikiprompt.org/wiki/deep-learning). He has also been involved in the European Laboratory for Learning and Intelligent Systems (ELLIS), chairing the network's board since its inception. ## Honors and Affiliations Schölkopf has received multiple awards, including the Royal Society Milner Award and the BBVA Foundation Frontiers of Knowledge Award (shared with [Vladimir Vapnik](https://www.wikiprompt.org/wiki/vladimir-vapnik) and Isabelle Guyon). He was elected a Fellow of the Royal Society in 2026. As of late 2023, he is an advisory board member of the French non-profit AI lab Kyutai, funded by Xavier Niel, Eric Schmidt, and others. His extensive publication record includes monographs on kernel methods and causal inference, and he remains one of the most cited researchers in computer science. His work has bridged theory and practice, influencing areas such as pattern recognition, neuroimaging, and climate science. ## External influence Schölkopf's ideas have shaped modern [artificial-intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) research. His kernel-based techniques, though partly superseded by [deep-learning](https://www.wikiprompt.org/wiki/deep-learning) in some domains, remain foundational for understanding high-dimensional data. His causal inference framework is increasingly integrated with [neural-network](https://www.wikiprompt.org/wiki/neural-network) models, enabling more robust and interpretable [machine-learning](https://www.wikiprompt.org/wiki/machine-learning) systems. He has also contributed to policy discussions on AI ethics and safety, participating in the IEEE Global Initiative on Ethically Aligned Design. --- Source: https://www.wikiprompt.org/wiki/bernhard-scholkopf License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/) Last updated: 2026-09-05T14:09:25.241934+00:00