Orchestrated Approximate Message Passing: A New Way of Information Integration from Multimodal Data

Event details
Date | 19.12.2023 |
Hour | 14:15 › 15:15 |
Speaker | Sagnik Nandy (UPenn) |
Location | |
Category | Conferences - Seminars |
Event Language | English |
Integrating information from multiple correlated datasets is a crucial aspect of modern data science. Despite the existence of numerous methods for this purpose, their statistical properties remain unclear, making valid statistical inference challenging. In this talk, I shall introduce Orchestrated Approximate Message Passing for integrating information across multiple correlated datasets. This method is both computationally efficient and statistically optimal under certain stylized models. Subsequently, I shall show how to apply this technique to tackle two distinct problems, namely community detection in a balanced, two-community, contextual stochastic block model and compressed sensing in the presence of network-side information.
Practical information
- Informed public
- Free
Organizer
- Lénaïc Chizat Lenka Zdeborova