Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

Local Semantics Refinement of Adaptive Representations for Robust Noisy Label Learning.

Yueer Lin, Yang Zhang, Jiao Hou

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society

|May 29, 2026

Summary

This summary is machine-generated.

Related Concept Videos

State Space Representation

State Space Representation

The frequency-domain technique, commonly used in analyzing and designing feedback control systems, is effective for linear, time-invariant systems. However, it falls short when dealing with nonlinear, time-varying, and multiple-input multiple-output systems. The time-domain or state-space approach addresses these limitations by utilizing state variables to construct simultaneous, first-order differential equations, known as state equations, for an nth-order system.
Consider an RLC circuit, a...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

CCR8 orchestrates an immunosuppressive niche in the liver to promote Echinococcus multilocularis infection.

PLoS neglected tropical diseases·2026

Same author

Hepatic macrophages play critical roles in the establishment and growth of hydatid cysts in the liver during Echinococcus granulosus sensu stricto infection.

PLoS neglected tropical diseases·2023

Same author

4'-Ethynyl-2'-deoxy-2'-β-fluoro-2-fluoroadenosine: A Highly Potent and Orally Available Clinical Candidate for the Treatment of HIV-1 Infection.

Journal of medicinal chemistry·2023

Same author

Activation of the phenylpropanoid pathway in Citrus sinensis collapsed vesicles during segment drying revealed by physicochemical and targeted metabolomics analysis.

Food chemistry·2023

Same author

Single-Cell Heterogeneity of the Liver-Infiltrating Lymphocytes in Individuals with Chronic Echinococcus multilocularis Infection.

Infection and immunity·2022

Same author

TIR-catalyzed ADP-ribosylation reactions produce signaling molecules for plant immunity.

Science (New York, N.Y.)·2022

Same journal

AgonicDreamer: Enhancing Multi-View Consistency in Text-to-3D Generation via Rectified Score Distillation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

BiCM-Prompt: Bidirectional Cross-Modal Prompt Tuning for Class-Incremental Learning on Multisource Remote Sensing Images.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

GoP-based Quality Enhancement on Video Compression.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Align then Tensorize: Multi-Level Consistent Anchor Graph Learning for Scalable Multi-View Clustering.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Beyond Fidelity: Diverse Image Synthesis via Retrieval-Augmented Diffusion.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Multi-Branch Tree-based Fusion Neural Architecture Search with Zero-Cost Screen for Multi-Modal Classification.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

See all related articles

Deep learning models struggle with noisy labels. Our new framework, Local Semantics Refinement of Adaptive Representations (LFDA), improves label quality by refining representations and aligning features, outperforming existing methods.

Area of Science:

Artificial Intelligence
Machine Learning
Computer Vision

Background:

Deep learning models require high-quality labeled data, but large datasets often contain inevitable noisy labels.
Current methods for handling noisy labels may exhibit confirmation bias and neglect valuable hard-but-clean samples.

Purpose of the Study:

To introduce a novel framework, Local Semantics Refinement of Adaptive Representations (LFDA), designed to adaptively refine label quality in deep learning datasets.
To address the limitations of existing label noise learning methods by leveraging local feature consistency and representation alignment.

Main Methods:

LFDA utilizes a Local Consistency Score module to assess sample similarity in latent space, distinguishing clean from noisy labels.
A confidence neighborhood is constructed for local reference guidance, improving the identification and correction of noisy instances.

Related Experiment Videos

The Reliability-Aware Representation Alignment (RRA) module aligns high-confidence representations to implicitly refine low-confidence instances via soft supervision.

Main Results:

LFDA consistently outperforms state-of-the-art methods on both synthetic and real-world noisy datasets.
Experimental results demonstrate LFDA's robustness and generalization ability across diverse noise conditions.
The framework effectively handles complex noise scenarios in label noise learning.

Conclusions:

LFDA offers a robust and effective solution for learning with noisy labels in deep learning.
The proposed method enhances label quality by leveraging local semantic information and adaptive representation refinement.
LFDA represents a significant advancement in addressing the challenge of noisy labels in large-scale datasets.