Clara Hoffmann and Prof. Nadja Klein from the SCC research group Methods for Big Data at the Karlsruhe Institute of Technology (KIT) have been awarded the prestigious Mitchell Prize sponsored by the Section on Bayesian Statistical Science (SBSS) of the ASA, the International Society for Bayesian Analysis (ISBA), and the Mitchell Prize Founders’ Committee.
Their paper, Marginally calibrated response distributions for end-to-end learning in autonomous driving, published in The Annals of Applied Statistics, was recognized for its outstanding contribution to statistical methodology with strong practical relevance. The research addresses uncertainty quantification in neural networks for autonomous driving – a key factor in ensuring the safety of such systems.
End-to-end models predict the steering angle directly from camera images but typically provide only point estimates. The method presented in the paper is based on the Implicit Copula Neural Linear Model (IC-NLM), which generates probability distributions for steering angles and is particularly well-calibrated. To make the approach scalable to large datasets, the authors employed a fast estimation technique using variational inference.
Tests with real driving data show that the IC-NLM is not only accurate but also more reliable in assessing its own predictive uncertainty compared to other models. This contributes to the explainability of neural networks. The award highlights the importance of statistical methods for safe AI applications – a key research focus at SCC.
Contact at SCC: Clara Hoffmann, Prof. Dr. Nadja Klein