Set a Thief to Catch a Thief: Combating Label Noise through Noisy Meta Learning

2025-02-22

AI Summary

This paper addresses learning with noisy labels in the absence of a clean validation set. We propose STCT, a meta-learning framework that requires no additional clean data. Its core innovation is the “noise-cleans-noise” paradigm: it treats noise-distributed training data itself as a meta-validation set, enabling decoupled optimization of noise-aware meta-correction and semi-supervised representation learning via alternating updates. STCT integrates noisy meta-learning, dynamic label refinement, consistency regularization, and semi-supervised learning. On CIFAR-10 under 80% symmetric label noise, it achieves 96.9% label correction accuracy and 95.2% classification accuracy—surpassing state-of-the-art methods. To our knowledge, STCT is the first approach to achieve highly robust noise label correction and model training without relying on any clean validation data.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Learning to Search
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Security and Privacy: Large-scale security measurements

Abstract

Learning from noisy labels (LNL) aims to train high-performance deep models using noisy datasets. Meta learning based label correction methods have demonstrated remarkable performance in LNL by designing various meta label rectification tasks. However, extra clean validation set is a prerequisite for these methods to perform label correction, requiring extra labor and greatly limiting their practicality. To tackle this issue, we propose a novel noisy meta label correction framework STCT, which counterintuitively uses noisy data to correct label noise, borrowing the spirit in the saying ``Set a Thief to Catch a Thief''. The core idea of STCT is to leverage noisy data which is i.i.d. with the training data as a validation set to evaluate model performance and perform label correction in a meta learning framework, eliminating the need for extra clean data. By decoupling the complex bi-level optimization in meta learning into representation learning and label correction, STCT is solved through an alternating training strategy between noisy meta correction and semi-supervised representation learning. Extensive experiments on synthetic and real-world datasets demonstrate the outstanding performance of STCT, particularly in high noise rate scenarios. STCT achieves 96.9% label correction and 95.2% classification performance on CIFAR-10 with 80% symmetric noise, significantly surpassing the current state-of-the-art.

Problem

• Eliminates need for clean validation data
• Corrects noisy labels using noisy data
• Improves performance in high noise scenarios

Innovation

• Noisy data for label correction
• Decoupled bi-level optimization strategy
• Alternating training for meta correction