Partially Observed Exchangeable Modeling
This addresses the problem of leveraging multiple related instances for improved dependency modeling in machine learning, offering a novel framework with broad applicability.
The paper tackles the problem of modeling dependencies among features across multiple related instances, proposing a framework that jointly models intra- and inter-instance dependencies to infer conditional distributions for unobserved dimensions. The result is a general framework achieving state-of-the-art performance across various applications.
Modeling dependencies among features is fundamental for many machine learning tasks. Although there are often multiple related instances that may be leveraged to inform conditional dependencies, typical approaches only model conditional dependencies over individual instances. In this work, we propose a novel framework, partially observed exchangeable modeling (POEx) that takes in a set of related partially observed instances and infers the conditional distribution for the unobserved dimensions over multiple elements. Our approach jointly models the intra-instance (among features in a point) and inter-instance (among multiple points in a set) dependencies in data. POEx is a general framework that encompasses many existing tasks such as point cloud expansion and few-shot generation, as well as new tasks like few-shot imputation. Despite its generality, extensive empirical evaluations show that our model achieves state-of-the-art performance across a range of applications.