ICNLSP 2023: Handling Realistic Label Noise in BERT Text Classification (Maha Agro, Hanan Aldarmaki)

Title of the presentation: Handling Realistic Label Noise in BERT Text Classification.
By: Maha Tufail Agro and Hanan Aldarmaki, MBZUAI, UAE.
6th International Conference on Natural Language and Speech Processing.
icnlsp.org/2023welcome
Abstract:
Label noise refers to errors in training labels caused by cheap data annotation methods, such
as web scraping or crowd-sourcing, which can be detrimental to the performance of supervised classifiers. Several methods have been proposed to counteract the effect of random label noise in supervised classification, and some studies have shown that BERT is already robust
against high rates of randomly injected label noise. However, real label noise is not random;
rather, it is often correlated with input features or other annotator-specific factors. In this paper, we evaluate BERT in the presence of two types of realistic label noise: feature-dependent
label noise, and synthetic label noise from annotator disagreements. We show that the presence of these types of noise significantly degrades BERT classification performance. To improve robustness, we evaluate different types of ensembles and noise-cleaning methods and
compare their effectiveness against label noise across different datasets.