Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

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

Low-Coverage Whole-Genome Sequencing Identifies Loci Associated with Birth Weight in East Friesian × Hu Crossbred Sheep.

Animals : an open access journal from MDPI·2026
Same author

TREM2 defends the liver against alveolar echinococcosis by driving fibrous shell formation.

Biochemical pharmacology·2026
Same author

Near-Infrared Polarity-Sensitive Fluorescent Probe for Distinguishing Normal and Cancer Cells and Imaging Mitochondrial Polarity Changes during Drug-Induced Liver Injury.

Analytical chemistry·2026
Same author

De Novo Designed Minibinders Targeting the GDF15-GFRAL Axis Reverse Cancer Cachexia and Restore Anti-Tumor Immunity.

Advanced science (Weinheim, Baden-Wurttemberg, Germany)·2026
Same author

Qingxuan Zhike Granules Modulate Gut Dysbiosis and Enhance Intestinal Repair in Murine <i>Mycoplasma pneumonia</i>.

Journal of microbiology and biotechnology·2026
Same author

The HIF-1α-Th17/Treg axis in pediatric infectious mononucleosis: a clinical investigation of metabolic-immune dysregulation.

BMC infectious diseases·2026

Related Experiment Video

Updated: Aug 20, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

663

MAGE: Multi-scale Context-aware Interaction based on Multi-granularity Embedding for Chinese Medical Question Answer

Meiling Wang1, Xiaohai He1, Yan Liu2

  • 1College of Electronics and Information Engineering, Sichuan University, Chengdu, Sichuan, 610065, China.

Computer Methods and Programs in Biomedicine
|November 24, 2022
PubMed
Summary

This study introduces a novel approach for Chinese medical question answer matching, enhancing accuracy by incorporating character glyphs and pinyin. The MAGE model significantly improves performance on key datasets, offering a potential tool for medical AI systems.

Keywords:
Attention mechanismMulti-granularity embeddingMulti-scale context-aware interactionQuestion answer matching

More Related Videos

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.6K
Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
05:47

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

Published on: June 13, 2025

502

Related Experiment Videos

Last Updated: Aug 20, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

663
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.6K
Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
05:47

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

Published on: June 13, 2025

502

Area of Science:

  • Natural Language Processing
  • Artificial Intelligence in Medicine
  • Computational Linguistics

Background:

  • The Chinese medical question answer matching (cMedQAM) task is crucial for medical question answering systems.
  • Existing deep learning and attention-based methods often overlook Chinese character specifics like glyphs and pinyin.
  • Current approaches may lose local semantic information by focusing solely on medical keywords.

Purpose of the Study:

  • To propose a multi-scale context-aware interaction approach based on multi-granularity embedding (MAGE).
  • To address the limitations of existing methods in capturing Chinese character features and local semantic context.
  • To improve the accuracy and effectiveness of Chinese medical question answer matching.

Main Methods:

  • Adapted ChineseBERT to integrate Chinese character glyphs and pinyin information, addressing homonyms.
  • Developed a context-aware interactive module for aligning question-answer sequences and inferring semantic relationships.
  • Employed a multi-view fusion method to combine local semantic features with attention representations.

Main Results:

  • Validated the MAGE approach on three public datasets: cMedQA V1.0, cMedQA V2.0, and cEpilepsyQA.
  • Achieved significant improvements in top-1 accuracy: 74.1% on cMedQA V1.0, 82.7% on cMedQA V2.0, and 60.9% on cEpilepsyQA.
  • Demonstrated superior performance compared to state-of-the-art methods for cMedQAM tasks.

Conclusions:

  • The proposed MAGE model effectively enhances the accuracy of Chinese medical question answer matching.
  • The model shows potential as an intelligent assistant for future Chinese medical question answering systems.