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 Video

Updated: Apr 30, 2026

Author Spotlight: Understanding Processing of Olfactory and Spatial Information by Brain with Real-Time Behavioral Analysis

Author Spotlight: Understanding Processing of Olfactory and Spatial Information by Brain with Real-Time Behavioral Analysis

Published on: September 20, 2024

MDFE-Net: A Meta-Learning Driven Dual-Branch Feature Extraction Network for E-Nose Sensor Drift Adaptation.

Qilong Yang¹, Jinxia Liu^1,2, Yan Shi¹

¹School of Automation Engineering, Northeast Electric Power University, Jilin 132012, China.

|April 28, 2026

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

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

Sort by

Same author

SIRT5 desuccinylates ME2 to suppress oxidative stress in bovine mammary epithelial cells.

Free radical biology & medicine·2026

Same author

Enriching microbes capable of fluorotelomer acid defluorination: thermodynamic constraints and experimental challenges.

NPJ emerging contaminants·2026

Same author

EEG Cross-Subject Taste Classification Method: A Meta-Learning Wavelet Graph Convolutional Neural Network Under Sweet and Bitter Stimuli.

Biosensors·2026

Same author

Diverse origins and chemical complexity of nanoplastics.

Nature nanotechnology·2026

Same author

Spatial Optimization of Electrophysiological Signal Acquisition in <i>Clivia</i> Leaves Under a Controlled Leaf-Surface Salt-Treatment Model.

Plants (Basel, Switzerland)·2026

Same author

HO-1 enhances autophagy to alleviate oxidative damage in BMECs via Keap1/Nrf2 and AMPK/mTOR signaling.

Cellular signalling·2026

Same journal

Plant-Plant Communication for Systemic Acquired Resistance under Biotic Stress Spatiotemporally Tracked by an <i>In Situ</i> Surface-Enhanced Raman Spectroscopy Aerosol Spraying Analyzer.

ACS sensors·2026

Same journal

Modulating Electronic Structure via Bimetallic D<i>-</i>Band Engineering toward an Ultrasensitive Sensor Platform for Caffeic Acid in Food.

ACS sensors·2026

Same journal

Indiscriminate <i>T</i><i>rans</i>-Cleavage Activity of CRISPR/SuCas12a2 Enables Sensitive Detection of SARS-CoV-2.

ACS sensors·2026

Same journal

Spin-State Engineering in 2D Metal-Organic Frameworks for Ultrasensitive Room-Temperature Ammonia Sensing.

ACS sensors·2026

Same journal

A Wearable Microneedle-Based Electrochemical Aptamer Sensor: Enabling Real-Time Dynamic NT-proBNP Monitoring for Enhanced Heart Failure Management.

ACS sensors·2026

Same journal

Double-Strand Gated Biosensor for Ultrasensitive T4 PNK Detection via λ-Exonuclease-Driven Background Suppression and Dimer G-Triplex Signal Amplification.

ACS sensors·2026

See all related articles

This study introduces a novel network for electronic nose (E-nose) systems to combat sensor drift. The developed MDFE-Net effectively compensates for drift using meta-learning, improving gas recognition accuracy with limited data.

Area of Science:

Sensor Technology
Artificial Intelligence
Machine Learning

Background:

Gas sensors suffer from ageing and environmental changes, leading to performance degradation in electronic nose (E-nose) systems.
Sensor drift significantly impacts the reliability and accuracy of E-nose applications over time.
Existing methods struggle with long-term drift and device variability, necessitating adaptive solutions.

Purpose of the Study:

To develop a Meta-learning Driven Dual-branch Feature Extraction E-nose Drift Adaptation Network (MDFE-Net) for few-shot drift compensation.
To enhance the robustness and generalization capabilities of E-nose systems against sensor drift.
To enable rapid adaptation to new drift conditions using minimal labeled data.

Main Methods:

A dual-branch feature extraction module (DBFE) was designed to capture temporal and spatial sensor response patterns.

Keywords:

E-nose few-shot adaptation meta-learning sensor drift triplet loss

Related Experiment Videos

Last Updated: Apr 30, 2026

Author Spotlight: Understanding Processing of Olfactory and Spatial Information by Brain with Real-Time Behavioral Analysis

Author Spotlight: Understanding Processing of Olfactory and Spatial Information by Brain with Real-Time Behavioral Analysis

Published on: September 20, 2024

Model-Agnostic Meta-Learning (MAML) was employed to learn transferable knowledge for quick adaptation.

Adaptive triplet loss and dynamically reweighted cross-entropy loss were utilized for improved feature discrimination.

Main Results:

MDFE-Net achieved high accuracy (95.53% long-term, 95.71% short-term drift) on the Gas Sensor Array Drift Dataset.
The network demonstrated strong cross-device generalization, reaching 97.89% accuracy on the Twin Gas Sensor Arrays benchmark.
The proposed method effectively mitigates long-term nonlinear drift and improves performance with scarce calibration labels.

Conclusions:

MDFE-Net successfully addresses the challenge of sensor drift in E-nose systems through a combination of deep metric learning and meta-learning.
The approach offers a robust solution for maintaining E-nose performance in real-world, long-term deployment scenarios.
This work paves the way for more reliable and adaptable E-nose applications with reduced calibration requirements.