Multi-Modal Conformal Prediction Regions with Simple Structures by Optimizing Convex Shape Templates

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tumu, Renukanandan, Cleaveland, Matthew, Mangharam, Rahul, Pappas, George J., Lindemann, Lars
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915182505099264
author Tumu, Renukanandan
Cleaveland, Matthew
Mangharam, Rahul
Pappas, George J.
Lindemann, Lars
author_facet Tumu, Renukanandan
Cleaveland, Matthew
Mangharam, Rahul
Pappas, George J.
Lindemann, Lars
contents Conformal prediction is a statistical tool for producing prediction regions for machine learning models that are valid with high probability. A key component of conformal prediction algorithms is a \emph{non-conformity score function} that quantifies how different a model's prediction is from the unknown ground truth value. Essentially, these functions determine the shape and the size of the conformal prediction regions. While prior work has gone into creating score functions that produce multi-model prediction regions, such regions are generally too complex for use in downstream planning and control problems. We propose a method that optimizes parameterized \emph{shape template functions} over calibration data, which results in non-conformity score functions that produce prediction regions with minimum volume. Our approach results in prediction regions that are \emph{multi-modal}, so they can properly capture residuals of distributions that have multiple modes, and \emph{practical}, so each region is convex and can be easily incorporated into downstream tasks, such as a motion planner using conformal prediction regions. Our method applies to general supervised learning tasks, while we illustrate its use in time-series prediction. We provide a toolbox and present illustrative case studies of F16 fighter jets and autonomous vehicles, showing an up to $68\%$ reduction in prediction region area compared to a circular baseline region.
format Preprint
id arxiv_https___arxiv_org_abs_2312_07434
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Multi-Modal Conformal Prediction Regions with Simple Structures by Optimizing Convex Shape Templates
Tumu, Renukanandan
Cleaveland, Matthew
Mangharam, Rahul
Pappas, George J.
Lindemann, Lars
Machine Learning
Systems and Control
Conformal prediction is a statistical tool for producing prediction regions for machine learning models that are valid with high probability. A key component of conformal prediction algorithms is a \emph{non-conformity score function} that quantifies how different a model's prediction is from the unknown ground truth value. Essentially, these functions determine the shape and the size of the conformal prediction regions. While prior work has gone into creating score functions that produce multi-model prediction regions, such regions are generally too complex for use in downstream planning and control problems. We propose a method that optimizes parameterized \emph{shape template functions} over calibration data, which results in non-conformity score functions that produce prediction regions with minimum volume. Our approach results in prediction regions that are \emph{multi-modal}, so they can properly capture residuals of distributions that have multiple modes, and \emph{practical}, so each region is convex and can be easily incorporated into downstream tasks, such as a motion planner using conformal prediction regions. Our method applies to general supervised learning tasks, while we illustrate its use in time-series prediction. We provide a toolbox and present illustrative case studies of F16 fighter jets and autonomous vehicles, showing an up to $68\%$ reduction in prediction region area compared to a circular baseline region.
title Multi-Modal Conformal Prediction Regions with Simple Structures by Optimizing Convex Shape Templates
topic Machine Learning
Systems and Control
url https://arxiv.org/abs/2312.07434