Why does AI call a Parrot a Sleeping Bag?
Presentation at the Young Researcher Symposium of RPTU at Fraunhofer in Kaiserslautern on training data biases in computer vision and how to measure and mitigate these.
slides
Presentations.
4 talks
OA-CutMix: Correcting the Label Bias of CutMix
Presentation at ICANN 2026 about how CutMix's label are deceiving in image classification and how we can efficiently correct for that.
Which Transformer to Favor: A Comparative Analysis of Efficiency in Vision Transformers
Presentation at WACV 2025 on a large-scale benchmark of 45+ transformer models for image classification, evaluating accuracy, speed, and memory efficiency.
TaylorShift: Shifting the Complexity of Self-Attention from Squared to Linear (and Back) using Taylor-Softmax
Oral presentation at ICPR 2024 introducing TaylorShift, a novel reformulation of the attention mechanism using Taylor-Softmax that enables full token-to-token interactions in linear time.