Likelihood Ratio Test
method
ID: likelihoodratiotest
Related Papers
Quantifying Privacy Risks of Masked Language Models Using Membership Inference Attacks
AbstractThe wide adoption and application of Masked language models (MLMs) on sensitive data (from legal to medical) necessitates a thorough quantitative investigation into their privacy vulnerabilities. Prior attempts at measuring …...