Partially Observed Exchangeable Modeling

TL;DR

This paper introduces POEx, a novel framework for modeling dependencies in sets of related, partially observed data, enabling improved inference of unobserved features across multiple instances.

Contribution

The paper proposes a general framework, POEx, that models both intra- and inter-instance dependencies in partially observed data, covering existing tasks and enabling new applications like few-shot imputation.

Findings

Achieves state-of-the-art performance across various tasks.

Effectively models dependencies among features and instances.

Encompasses existing methods and introduces new capabilities.

Abstract

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.