ChronoSelect: Robust Learning with Noisy Labels via Dynamics Temporal Memory

2025-07-24

AI Summary

To address the degradation of deep neural network generalization caused by label noise in real-world data, this paper proposes ChronoSelect—a novel framework centered on a dynamic temporal memory mechanism. This mechanism employs a four-stage sliding update architecture to compactly compress prediction histories, integrated with dual-branch consistency verification and dynamic trajectory analysis for adaptive tri-partitioning of samples into clean, noisy, and boundary categories. We provide theoretical guarantees proving the convergence and stability of the proposed mechanism under label noise. Extensive experiments on diverse synthetic and real-world noisy benchmarks—including CIFAR-10/100-N, WebVision, and Clothing1M—demonstrate that ChronoSelect consistently outperforms state-of-the-art methods, achieving new SOTA performance. The framework significantly enhances model robustness and generalization capability in the presence of label corruption.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Training deep neural networks on real-world datasets is often hampered by the presence of noisy labels, which can be memorized by over-parameterized models, leading to significant degradation in generalization performance. While existing methods for learning with noisy labels (LNL) have made considerable progress, they fundamentally suffer from static snapshot evaluations and fail to leverage the rich temporal dynamics of learning evolution. In this paper, we propose ChronoSelect (chrono denoting its temporal nature), a novel framework featuring an innovative four-stage memory architecture that compresses prediction history into compact temporal distributions. Our unique sliding update mechanism with controlled decay maintains only four dynamic memory units per sample, progressively emphasizing recent patterns while retaining essential historical knowledge. This enables precise three-way sample partitioning into clean, boundary, and noisy subsets through temporal trajectory analysis and dual-branch consistency. Theoretical guarantees prove the mechanism's convergence and stability under noisy conditions. Extensive experiments demonstrate ChronoSelect's state-of-the-art performance across synthetic and real-world benchmarks.

Problem

• Addressing noisy labels in deep learning datasets
• Overcoming static evaluation limitations in noisy label learning
• Enhancing generalization via temporal dynamics in training

Innovation

• Four-stage memory architecture compresses prediction history
• Sliding update mechanism with controlled decay
• Temporal trajectory analysis for sample partitioning