Meta Self-Refinement for Robust Learning with Weak Supervision

TL;DR

This paper introduces Meta Self-Refinement (MSR), a novel framework that enhances robustness of deep neural networks trained with weak supervision by refining pseudo-labels through meta-learning, leading to improved generalization and noise resistance.

Contribution

MSR is the first method to dynamically refine pseudo-labels via meta-gradient descent, significantly reducing error propagation from noisy weak labels.

Findings

MSR outperforms state-of-the-art methods by up to 11.4% in accuracy.

MSR demonstrates robustness against label noise across eight NLP benchmarks.

MSR effectively refines pseudo-labels to improve model generalization.

Abstract

Training deep neural networks (DNNs) under weak supervision has attracted increasing research attention as it can significantly reduce the annotation cost. However, labels from weak supervision can be noisy, and the high capacity of DNNs enables them to easily overfit the label noise, resulting in poor generalization. Recent methods leverage self-training to build noise-resistant models, in which a teacher trained under weak supervision is used to provide highly confident labels for teaching the students. Nevertheless, the teacher derived from such frameworks may have fitted a substantial amount of noise and therefore produce incorrect pseudo-labels with high confidence, leading to severe error propagation. In this work, we propose Meta Self-Refinement (MSR), a noise-resistant learning framework, to effectively combat label noise from weak supervision. Instead of relying on a fixed teacher trained with noisy labels, we encourage the teacher to refine its pseudo-labels. At each training step, MSR performs a meta gradient descent on the current mini-batch to maximize the student performance on a clean validation set. Extensive experimentation on eight NLP benchmarks demonstrates that MSR is robust against label noise in all settings and outperforms state-of-the-art methods by up to 11.4% in accuracy and 9.26% in F1 score.