Sebastian Lapuschkin
Head Of Explainable AI Group At Fraunhofer Heinrich Hertz Institute, HHI at World Conference on eXplainable Artificial Intelligence
Berlin, Germany
Summary
🤩
Rockstar🎓
Top SchoolSebastian Lapuschkin is a research leader and Head of the Explainable AI Group at Fraunhofer HHI, combining over a decade of hands-on ML research with strategic group leadership of ~70 contributors to push XAI from interpretation toward actionable model and data improvement. He earned a PhD with distinction from TU Berlin for pioneering work on Layer-wise Relevance Propagation and Deep Taylor Decomposition, and his work has been recognized with awards such as the Hugo-Geiger-Prize and a Pattern Recognition Best Paper Award. Sebastian drives reproducible, open-source XAI tooling—contributing to projects like iNNvestigate and providing PyTorch-focused libraries such as Zennit and LXT—while also developing meta-analyses and automated detectors for spurious “Clever Hans” model behavior. Equally at home in code and theory, he focuses on making explanations human-understandable and usable as feedback to improve models and datasets, with additional interests in efficient ML and visual analytics. Based in Berlin, he also shapes the XAI community via conference leadership and cross-institutional collaborations.
9 years of coding experience
7 years of employment as a software developer
PhD (Dr. rer. nat.), Machine Learning, PhD (Dr. rer. nat.), Machine Learning at Technische Universität Berlin
German, English
Github Skills (9)
package-manager10
package-management10
python10
package10
mnist7
imagenet7
keras6
deep-learning5
machine-learning5
Programming languages (3)
Github contributions (5)
albermax/innvestigate
Feb 2018 - Oct 2018
A toolbox to iNNvestigate neural networks' predictions!
Role in this project:
Back-end Developer
Back-end DeveloperContributions:235 commits, 2 PRs, 7 pushes in 8 months
Contributions summary:Sebastian's contributions primarily involve adding and modifying dependencies within the `setup.py` file, indicating a focus on project build and dependency management. The commits show additions of required packages such as "future", "matplotlib", "h5py", and "pillow". Furthermore, the user contributed to the example scripts, preparing a test file for LRP code and migrating all methods for LRP test plotting for both MNIST and ImageNet datasets.
pytorchdeep-learningtoolboxneural-networksmachine-learning
sebastian-lapuschkin/explaining-deep-clinical-gait-classification
Aug 2019 - Dec 2021
Code and Data used for the paper "Explaining Machine Learning Models for Clinical Gait Analysis"
Contributions:179 commits, 2 PRs, 44 pushes in 2 years 4 months
gaitclinicalgait-analysismachine-learningmachine-learning-models
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.