# Probabilistic Decoupling of Labels in Classification

## TL;DR

This paper introduces a probabilistic decoupling approach for labels in classification tasks, enabling flexible inference across various supervised and semi-supervised learning scenarios, including noisy and partial labels.

## Contribution

It proposes a novel probabilistic decoupling method that generalizes classification inference to handle diverse label conditions and learning paradigms.

## Key findings

- Effective on Fashion MNIST with simulated noise
- Performs well on 20 News Groups with partial labels
- Unified approach for multiple label noise scenarios

## Abstract

We investigate probabilistic decoupling of labels supplied for training, from the underlying classes for prediction. Decoupling enables an inference scheme general enough to implement many classification problems, including supervised, semi-supervised, positive-unlabelled, noisy-label and suggests a general solution to the multi-positive-unlabelled learning problem. We test the method on the Fashion MNIST and 20 News Groups datasets for performance benchmarks, where we simulate noise, partial labelling etc.