This page contains automatically translated content.
Back
Accepted article for the Workshop on Interactive Adaptive Learning (IAL), ECML PKDD 2022
The paper titled"A Practical Evaluation of Active Learning Approaches for Object Detection" Jan Schneegans, Maarten Bieshaar and Bernhard Sick examines a wide range of
of active learning selection criteria for the task of object detection. They also distinguish between approaches that select samples based on aleatory or epistemic uncertainty, and discuss the implicit mechanisms related to the active learning approaches that seem to influence the model performance.