Confident Learning: Estimating Uncertainty in Dataset Labels

TL;DR

Confident learning (CL) is a novel framework for estimating and improving label quality in datasets by identifying label errors, which enhances model training and data understanding across various data modalities.

Contribution

This paper introduces a generalized, provably consistent confident learning method that estimates label noise and errors, outperforming recent approaches and applicable to multiple data types.

Findings

CL accurately finds label errors under certain conditions.

CL improves model accuracy by cleaning datasets before training.

CL quantifies label overlap and mislabeling in large datasets.

Abstract

Learning exists in the context of data, yet notions of confidence typically focus on model predictions, not label quality. Confident learning (CL) is an alternative approach which focuses instead on label quality by characterizing and identifying label errors in datasets, based on the principles of pruning noisy data, counting with probabilistic thresholds to estimate noise, and ranking examples to train with confidence. Whereas numerous studies have developed these principles independently, here, we combine them, building on the assumption of a class-conditional noise process to directly estimate the joint distribution between noisy (given) labels and uncorrupted (unknown) labels. This results in a generalized CL which is provably consistent and experimentally performant. We present sufficient conditions where CL exactly finds label errors, and show CL performance exceeding seven recent competitive approaches for learning with noisy labels on the CIFAR dataset. Uniquely, the CL framework is not coupled to a specific data modality or model (e.g., we use CL to find several label errors in the presumed error-free MNIST dataset and improve sentiment classification on text data in Amazon Reviews). We also employ CL on ImageNet to quantify ontological class overlap (e.g., estimating 645 "missile" images are mislabeled as their parent class "projectile"), and moderately increase model accuracy (e.g., for ResNet) by cleaning data prior to training. These results are replicable using the open-source cleanlab release.