New method tunes privacy noise to keep AI explanations accurate

Original: Pushing the (Decision) Boundaries: Dynamically Calibrating Differentially Private Noise to Explainability in Federated Learning

Source: arxiv.org ↗