Defending Our Privacy With Backdoors

The proliferation of large AI models trained on uncurated, often sensitive web-scraped data has raised significant privacy concerns. One of the concerns is that adversaries can extract information about the training data using privacy attacks. Unfortunately, the task of removing specific information from the models without sacrificing performance is not straightforward and has proven to be challenging. We propose a rather easy yet effective defense based on backdoor attacks to remove private information such as names and faces of individuals …

adversaries ai models attacks backdoors can cs.cl cs.cr cs.cv cs.lg data defending extract information large performance privacy privacy concerns proliferation scraped sensitive task training training data web