Hello! I am Aparna, a Research Scientist with Privacy Preserving Machine Learning team which establishes state-of-the-art scientific advancements for Meta’s products and systems. I did my PhD at Purdue University, where I worked on developing Decentralized Federated Learning algorithms for heterogeneous data distributions with Professor Kaushik Roy. In addition, I also explored efficient decentralized learning algorithms designed to reduce computational, memory, and communication overhead. My current research interests include on-device learning, model personalization and private evolution.
You can access the list of my publications under the research page and my resume under the resume page.
Posts
-
Stochastic Gradient Push (SGP)
-
Decentralized Parallel Stochastic Gradient Descent (D-PSGD)
-
Gossip Averaging
-
Decentralized Learning
-
Introduction to Federated Learning
subscribe via RSS