11 papers
Empirical Evidence for Simply Connected Decision Regions in Image Classifiers
Arjhun Swaminathan, Mete Akgün
Understanding the topology of decision regions is central to explaining the inner workings of deep neural networks. Prior empirical work has provided evidence that these regions ar…
Scaling Unsupervised Multi-Source Federated Domain Adaptation through Group-Wise Discrepancy Minimization
Larissa Reichart, Cem Ata Baykara, Ali Burak Ãnal +2
Unsupervised multi-source domain adaptation (UMDA) leverages labeled data from multiple source domains to generalize to an unlabeled target. While federated UMDA addresses privacy…
Accelerating Targeted Hard-Label Adversarial Attacks in Low-Query Black-Box Settings
Arjhun Swaminathan, Mete Akgün
Deep neural networks for image classification remain vulnerable to adversarial examples -- small, imperceptible perturbations that induce misclassifications. In black-box settings,…
PP-GWAS: Privacy Preserving Multi-Site Genome-wide Association Studies
Arjhun Swaminathan, Anika Hannemann, Ali Burak Ãnal +2
Genome-wide association studies are pivotal in understanding the genetic underpinnings of complex traits and diseases. Collaborative, multi-site GWAS aim to enhance statistical pow…
Accurate and Private Diagnosis of Rare Genetic Syndromes from Facial Images with Federated Deep Learning
Ali Burak Ãnal, Cem Ata Baykara, Peter Krawitz +1
Machine learning has shown promise in facial dysmorphology, where characteristic facial features provide diagnostic clues for rare genetic disorders. GestaltMatcher, a leading fram…
Federated Learning for Epileptic Seizure Prediction Across Heterogeneous EEG Datasets
Cem Ata Baykara, Saurav Raj Pandey, Ali Burak Ãnal +2
Developing accurate and generalizable epileptic seizure prediction models from electroencephalography (EEG) data across multiple clinical sites is hindered by patient privacy regul…