1 paper · 1 filter
William Novak, Muhammad Abusaqer
Membership inference attacks (MIAs) try to determine whether a specific record was used to train a model, a privacy risk that matters in natural language processing (NLP), where tr…