Exchangeable Generative Models with Flow Scans

Authors

  • Christopher Bender The University of North Carolina
  • Kevin O'Connor The University of North Carolina
  • Yang Li The University of North Carolina
  • Juan Garcia The University of North Carolina
  • Junier Oliva The University of North Carolina
  • Manzil Zaheer Google

DOI:

https://doi.org/10.1609/aaai.v34i06.6562

Abstract

In this work, we develop a new approach to generative density estimation for exchangeable, non-i.i.d. data. The proposed framework, FlowScan, combines invertible flow transformations with a sorted scan to flexibly model the data while preserving exchangeability. Unlike most existing methods, FlowScan exploits the intradependencies within sets to learn both global and local structure. FlowScan represents the first approach that is able to apply sequential methods to exchangeable density estimation without resorting to averaging over all possible permutations. We achieve new state-of-the-art performance on point cloud and image set modeling.

Downloads

Published

2020-04-03

How to Cite

Bender, C., O’Connor, K., Li, Y., Garcia, J., Oliva, J., & Zaheer, M. (2020). Exchangeable Generative Models with Flow Scans. Proceedings of the AAAI Conference on Artificial Intelligence, 34(06), 10053-10060. https://doi.org/10.1609/aaai.v34i06.6562

Issue

Section

AAAI Technical Track: Reasoning under Uncertainty