Anchored Discrete Factor Analysis

TL;DR

This paper introduces a semi-supervised algorithm for learning discrete factor analysis models with arbitrary structures, leveraging anchor variables to recover latent moments and improve robustness, demonstrated on real-world tasks.

Contribution

The paper proposes a novel semi-supervised learning algorithm that uses anchor variables to recover latent moments and enhances robustness through marginal polytope optimization.

Findings

Successfully applied to tag prediction on Stack Overflow questions.

Effective in medical diagnosis in emergency department settings.

Improves robustness of method-of-moments algorithms.

Abstract

We present a semi-supervised learning algorithm for learning discrete factor analysis models with arbitrary structure on the latent variables. Our algorithm assumes that every latent variable has an "anchor", an observed variable with only that latent variable as its parent. Given such anchors, we show that it is possible to consistently recover moments of the latent variables and use these moments to learn complete models. We also introduce a new technique for improving the robustness of method-of-moment algorithms by optimizing over the marginal polytope or its relaxations. We evaluate our algorithm using two real-world tasks, tag prediction on questions from the Stack Overflow website and medical diagnosis in an emergency department.