Bayesian Latent Space Models for Graphs Are Misspecified: Toward Robust Inference via Generalized Posteriors
We prove that Bayesian Inference for Latent Space models on graphs is often misspecified and can become overconfident and poorly calibrated.
Graphs
Personalized mapping of body homeostasis using whole-body PET connectomics and routine FDG PET imaging
Our first work on developping fully individualized connectomes for body homeostasis using PET imaging.
Aligning the Unseen in Attributed Graphs: Interplay between Graph Geometry and Node Attributes Manifold
Using a VAE framework, we investigate the potential misalignment between graph geometry and attributes geometry in attributed graphs.
Generating social networks with static and dynamic utility-maximization approaches
This paper offers a novel theoretical framework at the interface of physics and economics to understand and generate social networks using a dot-product latent space.