Parth Papaiyawala

Master's student at Brandenburg University of Technology (BTU) Cottbus-Senftenberg under the supervision of Dr. Mahdi Taheri. My research focuses on the reliability and security of deep neural networks in federated learning systems. I study how hardware fault injection (bit-flip attacks) and backdoor attacks interact and amplify each other in both centralized and distributed learning environments. My work covers experiments on image classification (CIFAR-10/ResNet) and event-camera object detection (Gen1 dataset), with the goal of understanding and quantifying the combined vulnerability of neural networks under simultaneous security and reliability threats.