Training Beyond Convergence: Grokking nnU-Net for Glioma Segmentation in Sub-Saharan MRI

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Barakat, Mohtady, Salah, Omar, Yasser, Ahmed, Ahmed, Mostafa, Arief, Zahirul, Khan, Waleed, Zhang, Dong, Iorumbur, Aondona, Raymond, Confidence, Barakat, Mohannad, Magdy, Noha
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908799188598784
author Barakat, Mohtady
Salah, Omar
Yasser, Ahmed
Ahmed, Mostafa
Arief, Zahirul
Khan, Waleed
Zhang, Dong
Iorumbur, Aondona
Raymond, Confidence
Barakat, Mohannad
Magdy, Noha
author_facet Barakat, Mohtady
Salah, Omar
Yasser, Ahmed
Ahmed, Mostafa
Arief, Zahirul
Khan, Waleed
Zhang, Dong
Iorumbur, Aondona
Raymond, Confidence
Barakat, Mohannad
Magdy, Noha
contents Gliomas are placing an increasingly clinical burden on Sub-Saharan Africa (SSA). In the region, the median survival for patients remains under two years, and access to diagnostic imaging is extremely limited. These constraints highlight an urgent need for automated tools that can extract the maximum possible information from each available scan, tools that are specifically trained on local data, rather than adapted from high-income settings where conditions are vastly different. We utilize the Brain Tumor Segmentation (BraTS) Africa 2025 Challenge dataset, an expert annotated collection of glioma MRIs. Our objectives are: (i) establish a strong baseline with nnUNet on this dataset, and (ii) explore whether the celebrated "grokking" phenomenon an abrupt, late training jump from memorization to superior generalization can be triggered to push performance without extra labels. We evaluate two training regimes. The first is a fast, budget-conscious approach that limits optimization to just a few epochs, reflecting the constrained GPU resources typically available in African institutions. Despite this limitation, nnUNet achieves strong Dice scores: 92.3% for whole tumor (WH), 86.6% for tumor core (TC), and 86.3% for enhancing tumor (ET). The second regime extends training well beyond the point of convergence, aiming to trigger a grokking-driven performance leap. With this approach, we were able to achieve grokking and enhanced our results to higher Dice scores: 92.2% for whole tumor (WH), 90.1% for tumor core (TC), and 90.2% for enhancing tumor (ET).
format Preprint
id arxiv_https___arxiv_org_abs_2601_22637
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Training Beyond Convergence: Grokking nnU-Net for Glioma Segmentation in Sub-Saharan MRI
Barakat, Mohtady
Salah, Omar
Yasser, Ahmed
Ahmed, Mostafa
Arief, Zahirul
Khan, Waleed
Zhang, Dong
Iorumbur, Aondona
Raymond, Confidence
Barakat, Mohannad
Magdy, Noha
Image and Video Processing
Artificial Intelligence
Computer Vision and Pattern Recognition
Gliomas are placing an increasingly clinical burden on Sub-Saharan Africa (SSA). In the region, the median survival for patients remains under two years, and access to diagnostic imaging is extremely limited. These constraints highlight an urgent need for automated tools that can extract the maximum possible information from each available scan, tools that are specifically trained on local data, rather than adapted from high-income settings where conditions are vastly different. We utilize the Brain Tumor Segmentation (BraTS) Africa 2025 Challenge dataset, an expert annotated collection of glioma MRIs. Our objectives are: (i) establish a strong baseline with nnUNet on this dataset, and (ii) explore whether the celebrated "grokking" phenomenon an abrupt, late training jump from memorization to superior generalization can be triggered to push performance without extra labels. We evaluate two training regimes. The first is a fast, budget-conscious approach that limits optimization to just a few epochs, reflecting the constrained GPU resources typically available in African institutions. Despite this limitation, nnUNet achieves strong Dice scores: 92.3% for whole tumor (WH), 86.6% for tumor core (TC), and 86.3% for enhancing tumor (ET). The second regime extends training well beyond the point of convergence, aiming to trigger a grokking-driven performance leap. With this approach, we were able to achieve grokking and enhanced our results to higher Dice scores: 92.2% for whole tumor (WH), 90.1% for tumor core (TC), and 90.2% for enhancing tumor (ET).
title Training Beyond Convergence: Grokking nnU-Net for Glioma Segmentation in Sub-Saharan MRI
topic Image and Video Processing
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.22637