https://compai-lab.io/author/kiechel/ kiechel Wowchemy (https://wowchemy.com)en-usThu, 01 Oct 2026 00:00:00 +0000 https://compai-lab.io/media/icon_hu790efcb2e4090d1e7a0ffec0a0776e8f_331139_512x512_fill_lanczos_center_3.png https://compai-lab.io/author/kiechel/ https://compai-lab.io/post/26_10_neurips_paper/ Thu, 01 Oct 2026 00:00:00 +0000 https://compai-lab.io/post/26_10_neurips_paper/ <p>Our lab will be presenting one paper at the The Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS) 2026 as a Spotlight:</p> <ul> <li><strong>Entropy Minimization without Model Collapse: Mitigating Prediction Bias for Medical Imaging</strong><br> Tim Nielen, Sameer Ambekar, Johannes Kiechle, Daniel M. Lang, Julia A. Schnabel<br> (<a href="https://arxiv.org/pdf/2606.02339" target="_blank" rel="noopener">https://arxiv.org/pdf/2606.02339</a>)<br/><br/></li> </ul> <p>TL;DR: We identify the root cause of model collapse: Entropy Minimization, which amplifies prediction bias and ultimately drives models into collapse. Building on this insight, we propose Distribution Shift Bias Reduction (DSBR), a method that effectively mitigates model collapse across both natural vision and medical imaging settings.</p> <p><br/><br/></p>