Published in World Conference on Explainable Artificial Intelligence (xAI 2025), 2025
Proposed a biologically inspired method for assessing convolutional filter significance using SSVEP-based frequency tagging to enhance model interpretability in CNNs.
Published in arXiv preprint, 2025
Introduced ROAR, a scalable privacy-preserving framework that removes sensitive objects using segmentation and generative inpainting while preserving dataset utility for 2D and 3D vision tasks.
Published in ACM Journal: Digital Threats: Research and Practice (DTRAP), 2024
Developed a hybrid deep learning approach for behavior-based threat actor attribution, employing natural language processing and machine learning.