Search papers, labs, and topics across Lattice.
This paper introduces a new class of estimators for the Sliced Wasserstein (SW) distance that utilize cumulative distribution functions (CDFs) instead of quantile functions, eliminating the need for sorting projected samples and enabling efficient scaling through data parallelism. The proposed estimators are particularly advantageous in scenarios where CDFs are more accessible than quantile functions, such as in mixtures of Gaussians, and they seamlessly integrate with federated learning frameworks by allowing local computation and aggregation of CDFs. Key results demonstrate that these estimators maintain competitive performance while significantly reducing computational overhead, making them a practical alternative for large-scale applications.
New estimators for the Sliced Wasserstein distance leverage cumulative distribution functions to enhance scalability and efficiency in large datasets and federated learning contexts.
The Sliced Wasserstein (SW) distance has emerged as a computationally attractive alternative to the Wasserstein distance by leveraging one-dimensional optimal transport along random projections. Standard estimators of the SW distance rely on Monte Carlo averages of one-dimensional Wasserstein distances computed via quantile functions, which require sorting projected samples and access to full datasets. In this work, we introduce a new class of estimators for the Sliced Wasserstein distance based on cumulative distribution functions (CDFs) of projected measures, that avoid sorting and scale via massive dataset parallelism. This class includes several estimators, some of them being indexed by hyperparameters controlling their variance or smoothness. We show that they are especially well suited to scenarios in which CDFs are more tractable than quantile functions, such as mixtures of Gaussians, and moreover that they are also naturally compatible with federated learning, since CDFs of projected data can be computed and aggregated locally without requiring the exchange of raw samples.
