Fine-tuning Vision Classifiers On A Budget

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kumar, Sunil, Sandler, Ted, Varshavskaya, Paulina
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913524818640896
author Kumar, Sunil
Sandler, Ted
Varshavskaya, Paulina
author_facet Kumar, Sunil
Sandler, Ted
Varshavskaya, Paulina
contents Fine-tuning modern computer vision models requires accurately labeled data for which the ground truth may not exist, but a set of multiple labels can be obtained from labelers of variable accuracy. We tie the notion of label quality to confidence in labeler accuracy and show that, when prior estimates of labeler accuracy are available, using a simple naive-Bayes model to estimate the true labels allows us to label more data on a fixed budget without compromising label or fine-tuning quality. We present experiments on a dataset of industrial images that demonstrates that our method, called Ground Truth Extension (GTX), enables fine-tuning ML models using fewer human labels.
format Preprint
id arxiv_https___arxiv_org_abs_2410_00085
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Fine-tuning Vision Classifiers On A Budget
Kumar, Sunil
Sandler, Ted
Varshavskaya, Paulina
Machine Learning
Computer Vision and Pattern Recognition
Fine-tuning modern computer vision models requires accurately labeled data for which the ground truth may not exist, but a set of multiple labels can be obtained from labelers of variable accuracy. We tie the notion of label quality to confidence in labeler accuracy and show that, when prior estimates of labeler accuracy are available, using a simple naive-Bayes model to estimate the true labels allows us to label more data on a fixed budget without compromising label or fine-tuning quality. We present experiments on a dataset of industrial images that demonstrates that our method, called Ground Truth Extension (GTX), enables fine-tuning ML models using fewer human labels.
title Fine-tuning Vision Classifiers On A Budget
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.00085