Learning in an Uncertain World: Representing Ambiguity Through Multiple Hypotheses

TL;DR

This paper introduces a framework for modeling uncertainty in prediction tasks by reformulating single-prediction models into multiple hypothesis prediction models, improving performance and revealing prediction variability across diverse applications.

Contribution

It presents a novel method to convert existing models into multi-hypothesis models with a new training procedure, applicable across various tasks.

Findings

MHP models outperform single-hypothesis models in all tested applications.

MHP models reveal valuable insights into prediction variability.

The approach is effective in human pose estimation, future prediction, image classification, and segmentation.

Abstract

Many prediction tasks contain uncertainty. In some cases, uncertainty is inherent in the task itself. In future prediction, for example, many distinct outcomes are equally valid. In other cases, uncertainty arises from the way data is labeled. For example, in object detection, many objects of interest often go unlabeled, and in human pose estimation, occluded joints are often labeled with ambiguous values. In this work we focus on a principled approach for handling such scenarios. In particular, we propose a framework for reformulating existing single-prediction models as multiple hypothesis prediction (MHP) models and an associated meta loss and optimization procedure to train them. To demonstrate our approach, we consider four diverse applications: human pose estimation, future prediction, image classification and segmentation. We find that MHP models outperform their single-hypothesis counterparts in all cases, and that MHP models simultaneously expose valuable insights into the variability of predictions.