Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only! Grab it Explore a technical lecture on differential privacy featuring Prof. Sheetal Kalyani from IIT Madras, who introduces a novel additive noise mechanism called Flipped Huber. Learn about this hybrid density approach that combines the advantages of Laplace and Gaussian distributions, offering both a sharp center and light sub-Gaussian tail characteristics. Discover the theoretical analysis behind this mechanism, including necessary and sufficient conditions for approximate differential privacy in one dimension and sufficient conditions in higher dimensions. Compare the effectiveness of this new mechanism through numerical simulations that demonstrate improved trade-offs between privacy and accuracy compared to existing methods. Gain insights from Prof. Kalyani's extensive expertise in wireless technology, differential privacy, extreme value theory, and machine learning applications in wireless systems.