Lorenz Linhardt

Lorenz Linhardt is a Researcher and PhD candidate at the Machine Learning Group of Technical University of Berlin, affiliated with the Berlin Institute for the Foundations of Learning and Data (BIFOLD). His research focuses on robustness of deep neural networks, spurious correlations, and representation learning, with applications in explainable AI and medical domains.

Educational Background:

• M.Sc. in Computer Science, 2019 – ETH Zürich
• B.Sc. in Computer Science, 2016 – University of Vienna

Research Interests:

• Robust Machine Learning: Addressing vulnerabilities in neural networks caused by spurious correlations and adversarial examples.
• Human Alignment: Bridging gaps between AI outputs and human judgments via representation learning and similarity metrics.
• Medical Applications: Leveraging machine learning for counterfactual inference in healthcare and dose-response modeling.

Article Trends: Recent work emphasizes alignment between human judgments and AI systems, with studies on latent diffusion models, adversarial robustness, and forensic analysis of malware classifiers. Earlier contributions span astronomy surveys and decision tree optimization.

Labs/Teams: Active in BIFOLD and the TU Berlin Machine Learning Group, focusing on foundational AI research.