Interpretable Distribution Features with Maximum Testing Power

Abstract

Two semimetrics on probability distributions are proposed, given as the sum of differences of expectations of analytic functions evaluated at spatial or frequency locations (i.e, features). The features are chosen so as to maximize the distinguishability of the distributions, by optimizing a lower bound on test power for a statistical test using these features. The result is a parsimonious and interpretable indication of how and where two distributions differ locally. An empirical estimate of the test power criterion converges with increasing sample size, ensuring the quality of the returned features. In real-world benchmarks on high-dimensional text and image data, linear-time tests using the proposed semimetrics achieve comparable performance to the state-of-the-art quadratic-time maximum mean discrepancy test, while returning human-interpretable features that explain the test results.

Cited by in corpus (8)

• Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration
• A Linear-Time Kernel Goodness-of-Fit Test
• Approximating the Quadratic Transportation Metric in Near-Linear Time
• Methods for Quantifying Dataset Similarity: a Review, Taxonomy and Comparison
• DeepClimGAN: A High-Resolution Climate Data Generator
• Model Reuse with Reduced Kernel Mean Embedding Specification
• Testing and Learning on Distributions with Symmetric Noise Invariance
• Post Selection Inference with Incomplete Maximum Mean Discrepancy Estimator