Simone Stumpf

Papers

Total Citations

H-Index

About

Simone Stumpf is a leading researcher in human-centered AI, with key contributions spanning explainable AI (XAI), interactive machine learning, and teachable object recognition. Her work fundamentally explores how humans and AI systems can collaborate more effectively, particularly when non-experts need to understand, trust, and customize AI behavior. A major contribution is the development of the ORBIT dataset, a real-world few-shot benchmark that enables object recognition systems to learn new categories from just a handful of user-provided examples—a critical step toward practical personalization in robotics and assistive technologies. This dataset, published in 2021, has garnered significant attention (over 30 citations) for addressing the gap between data-hungry deep learning and real-world user needs. Stumpf is also widely recognized for pioneering research on interpretable machine learning, where she investigates how to present model explanations to end-users in ways that foster appropriate trust and enable meaningful feedback. Her work consistently bridges technical AI advances with rigorous user studies, ensuring that AI systems are not only powerful but also transparent and controllable by the people who use them.