Iterative Alignment Flows

TL;DR

This paper introduces an iterative, non-adversarial flow-based method for aligning multiple distributions efficiently in a shared latent space, leveraging optimal transport theory for improved performance.

Contribution

It proposes a novel iterative approach that jointly aligns multiple distributions without adversarial training, inspired by OT algorithms, enhancing efficiency and scalability.

Findings

Achieves competitive alignment quality with low computational cost.

Effectively handles more than two distributions simultaneously.

Outperforms some existing methods in empirical tests.

Abstract

The unsupervised task of aligning two or more distributions in a shared latent space has many applications including fair representations, batch effect mitigation, and unsupervised domain adaptation. Existing flow-based approaches estimate multiple flows independently, which is equivalent to learning multiple full generative models. Other approaches require adversarial learning, which can be computationally expensive and challenging to optimize. Thus, we aim to jointly align multiple distributions while avoiding adversarial learning. Inspired by efficient alignment algorithms from optimal transport (OT) theory for univariate distributions, we develop a simple iterative method to build deep and expressive flows. Our method decouples each iteration into two subproblems: 1) form a variational approximation of a distribution divergence and 2) minimize this variational approximation via closed-form invertible alignment maps based on known OT results. Our empirical results give evidence that this iterative algorithm achieves competitive distribution alignment at low computational cost while being able to naturally handle more than two distributions.