Binary Losses for Density Ratio Estimation

TL;DR

This paper investigates how the choice of binary loss functions affects density ratio estimation accuracy, providing a theoretical framework for designing loss functions that prioritize different density ratio ranges, and demonstrating improved performance in real-world tasks.

Contribution

It characterizes all loss functions that yield small error in density ratio estimation within a Bregman divergence framework and offers a recipe for constructing loss functions with desired properties.

Findings

Novel loss functions outperform existing methods in 484 real-world tasks.

Theoretical characterization of loss functions for density ratio estimation.

Improved parameter choice in deep domain adaptation algorithms.

Abstract

Estimating the ratio of two probability densities from a finite number of observations is a central machine learning problem. A common approach is to construct estimators using binary classifiers that distinguish observations from the two densities. However, the accuracy of these estimators depends on the choice of the binary loss function, raising the question of which loss function to choose based on desired error properties. For example, traditional loss functions, such as logistic or boosting loss, prioritize accurate estimation of small density ratio values over large ones, even though the latter are more critical in many applications. In this work, we start with prescribed error measures in a class of Bregman divergences and characterize all loss functions that result in density ratio estimators with small error. Our characterization extends results on composite binary losses from (Reid & Williamson, 2010) and their connection to density ratio estimation as identified by (Menon & Ong, 2016). As a result, we obtain a simple recipe for constructing loss functions with certain properties, such as those that prioritize an accurate estimation of large density ratio values. Our novel loss functions outperform related approaches for resolving parameter choice issues of 11 deep domain adaptation algorithms in average performance across 484 real-world tasks including sensor signals, texts, and images.