Oberseminar "Mathematik des Maschinellen Lernens und Angewandte Analysis" - Dr. Muni Sreenivas Pydi
Differentially Private Gradient Flow for Generative Modeling
Date: | 04/30/2024, 4:15 PM - 5:15 PM |
Category: | Veranstaltung |
Location: | Hubland Süd, Geb. Z6 (Zentrales Hörsaal- u. Seminargebäude), 0.002 |
Organizer: | Center for Artificial Intelligence and Data Science (CAIDAS) |
Speaker: | Dr. Muni Sreenivas Pydi, Universite PSL, Paris |
Safeguarding privacy in sensitive training data is paramount, particularly in the context of generative modeling. This is done through either differentially private stochastic gradient descent, or with a differentially private metric for training models or generators. In this talk, I will introduce a novel differentially private generative modeling approach based on parameter-free gradient flows in the space of probability measures. The proposed algorithm is a new discretized flow which operates through a particle scheme, utilizing drift derived from the sliced Wasserstein distance and computed in a private manner. Our experiments show that compared to a generator-based model, our proposed model can generate higher-fidelity data at a low privacy budget, offering a viable alternative to generator-based approaches.
Der Vortrag findet im Rahmen der AI Talks @ JMU statt.