Publications

Quantifying the Predictive Uncertainty of Regression GNN Models Under Target Domain Shifts

Published in MICCAI PRIME workshop, 2022

We propose a deep graph ensemble of regression GNNs for estimating predictive uncertainty under domain shifts.

Recommended citation: S. Yürekli, M. A. Demirtaş, and I. Rekik, “Quantifying the Predictive Uncertainty of Regression GNN Models Under Target Domain Shifts,” in Predictive Intelligence in Medicine, Cham, 2022, pp. 149–159. doi: 10.1007/978-3-031-16919-9_14. https://doi.org/10.1007/978-3-031-16919-9_14

Semantic Parsing of Interpage Relations

Accepted in ICPR 2022, 2022

We formalize the task of semantic parsing for parsing relations between pages in multi-page documents as dependency trees.

Recommended citation: M. A. Demirtas, B. Oral, M. Y. Akpinar, and O. Deniz, “Semantic Parsing of Interpage Relations,” in 2022 26th International Conference on Pattern Recognition (ICPR), Montreal, QC, Canada, Aug. 2022, pp. 1579–1585. doi: 10.1109/ICPR56361.2022.9956546. https://ieeexplore.ieee.org/document/9956546

Predicting cognitive scores with graph neural networks through sample selection learning

Published in Brain Imaging and Behavior, 2021

We design a novel regression GNN model (namely RegGNN) for predicting IQ scores from brain connectivity.

Recommended citation: M. Hanik, M. A. Demirtaş, M. A. Gharsallaoui, and I. Rekik, “Predicting cognitive scores with graph neural networks through sample selection learning,” Brain Imaging and Behavior, vol. 16, no. 3, pp. 1123–1138, Jun. 2022, doi: 10.1007/s11682-021-00585-7. https://doi.org/10.1007/s11682-021-00585-7