Identity Inference

Identity inference focuses on determining an individual's identity from various data sources, aiming to understand and mitigate the risks of privacy breaches. Current research explores this across diverse data modalities, including visual data (e.g., images, shadows), biometric data (e.g., fingerprints), and textual data used to train large language models. Methods range from traditional machine learning approaches to advanced techniques like generative adversarial networks and Bayesian methods for uncertainty quantification in deep learning models. This research is crucial for developing robust privacy-preserving technologies and for assessing the vulnerabilities of increasingly prevalent AI systems.