Neural approximation of Wasserstein distance via a universal architecture for symmetric and factorwise group invariant functions

TL;DR

This paper introduces a neural network architecture capable of approximating Wasserstein distances between point sets, invariant to group actions, with model complexity independent of input size, and demonstrates superior empirical performance.

Contribution

The paper presents a universal neural network architecture for symmetric and factor-wise group invariant functions, specifically approximating Wasserstein distance with size-independent complexity.

Findings

Neural network can approximate Wasserstein distance with bounded complexity.

The proposed model outperforms state-of-the-art methods in accuracy and training speed.

Model generalizes well to different geometric optimization tasks.

Abstract

Learning distance functions between complex objects, such as the Wasserstein distance to compare point sets, is a common goal in machine learning applications. However, functions on such complex objects (e.g., point sets and graphs) are often required to be invariant to a wide variety of group actions e.g. permutation or rigid transformation. Therefore, continuous and symmetric product functions (such as distance functions) on such complex objects must also be invariant to the product of such group actions. We call these functions symmetric and factor-wise group invariant (or SFGI functions in short). In this paper, we first present a general neural network architecture for approximating SFGI functions. The main contribution of this paper combines this general neural network with a sketching idea to develop a specific and efficient neural network which can approximate the $p$-th Wasserstein distance between point sets. Very importantly, the required model complexity is independent of the sizes of input point sets. On the theoretical front, to the best of our knowledge, this is the first result showing that there exists a neural network with the capacity to approximate Wasserstein distance with bounded model complexity. Our work provides an interesting integration of sketching ideas for geometric problems with universal approximation of symmetric functions. On the empirical front, we present a range of results showing that our newly proposed neural network architecture performs comparatively or better than other models (including a SOTA Siamese Autoencoder based approach). In particular, our neural network generalizes significantly better and trains much faster than the SOTA Siamese AE. Finally, this line of investigation could be useful in exploring effective neural network design for solving a broad range of geometric optimization problems (e.g., $k$-means in a metric space).