Günter Klambauer
Also known as Gunter Klambauer
Günter Klambauer combines machine learning research with life science applications. He studied mathematics and biology in Vienna, began research at JKU Linz in 2010 and earned his doctorate there in 2014. His work includes DeepTox for predicting toxicity and methods used in the Tox21 challenge. In 2017, he and three co-authors introduced self-normalizing neural networks, using scaled exponential linear units to control activation statistics across layers. His research programme studies deep learning methods and their use in molecular biology and drug discovery.
- Now at
- Johannes Kepler University of Linz
- Works
- 113
- Citations
- 9,092
- h-index
- 39
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
2010–2014
Biology and machine learning
After studying mathematics and biology in Vienna, Klambauer began research in bioinformatics and machine learning at JKU Linz in 2010. He earned his doctorate there in 2014. His early work focused on machine-learning methods for genetics.
-
2015
Predicting molecular toxicity
Klambauer contributed to winning methods in the 2015 Tox21 Data Challenge, with DeepTox as a key paper. This work applies deep learning to the prediction of molecular toxicity.
-
2017
Self-normalizing neural networks
Klambauer, Thomas Unterthiner, Andreas Mayr, and Sepp Hochreiter introduced self-normalizing neural networks in 2017. The method uses scaled exponential linear units, or SELUs. The paper proves convergence of activation statistics under stated conditions and reports tests on tabular, drug discovery, and astronomy tasks.
-
As of October 2026
Deep learning for drug discovery
He focuses on deep learning for drug discovery and leads an AI in Life Sciences group. He also works with pharmaceutical industry data analysis groups.
Sources and credits
Photo: Barbara Amon; JKU. License: Not stated by source. Source: jku.at. Cropped to a square and resized. Ask us to remove or replace this photo.