Probabilistic Formulations of Regression with Mixed Guidance

TL;DR

This paper introduces a probabilistic regression framework that effectively incorporates both strong and weak guidance, such as relative orderings and bounds, enabling more flexible and easier data annotation.

Contribution

It proposes a novel convex probabilistic formulation for regression that integrates weak guidance types, expanding beyond traditional strong annotations.

Findings

Improved regression accuracy with weak guidance.

Convex optimization problems for the proposed formulations.

Enhanced flexibility in data annotation processes.

Abstract

Regression problems assume every instance is annotated (labeled) with a real value, a form of annotation we call \emph{strong guidance}. In order for these annotations to be accurate, they must be the result of a precise experiment or measurement. However, in some cases additional \emph{weak guidance} might be given by imprecise measurements, a domain expert or even crowd sourcing. Current formulations of regression are unable to use both types of guidance. We propose a regression framework that can also incorporate weak guidance based on relative orderings, bounds, neighboring and similarity relations. Consider learning to predict ages from portrait images, these new types of guidance allow weaker forms of guidance such as stating a person is in their 20s or two people are similar in age. These types of annotations can be easier to generate than strong guidance. We introduce a probabilistic formulation for these forms of weak guidance and show that the resulting optimization problems are convex. Our experimental results show the benefits of these formulations on several data sets.