Related Experiment Video
Updated: Sep 24, 2025
13:18
Network Pharmacology Prediction and Experimental Validation of Trichosanthes-Fritillaria thunbergii Action Mechanism Against Lung Adenocarcinoma
Published on: March 3, 2023
1.4K
An Improved Multitask Learning Model with Matching Network and Its Application in Traditional Chinese Medicine
Yingshuai Wang1,2, Jing-Han Xu3, Meng Zhang1,2
1Department of Computer, School of Computer and Communication Engineering, University of Science and Technology Beijing (USTB), Beijing 100083, China.
Journal of Healthcare Engineering
|May 9, 2022
Summary
This study introduces a novel MMOE-match network for traditional Chinese medicine (TCM) recommendations, improving medical record accuracy. The new model integrates matching networks with multitask learning (MTL) for better syndrome element and prescription recommendations.
Area of Science:
- Artificial Intelligence
- Medical Informatics
- Traditional Chinese Medicine
Background:
- Multitask learning (MTL) is crucial for various applications but existing methods lack matching relation features.
- Accurate medical record recommendation is vital for intelligent medical treatment in Traditional Chinese Medicine (TCM).
Purpose of the Study:
- To propose a novel MMOE-match network for modeling matches between medical cases and syndrome elements.
- To introduce recommendation algorithms into TCM studies for improved medical record recommendation.
Main Methods:
- Developed a novel MMOE-match network combining a two-tower matching network and multitask learning.
- Integrated matching network outputs as input for multitask learning.
- Compared manually designed matching features with model-generated ones.
Main Results:
- The proposed MMOE-match network demonstrated significant positive benefits in TCM recommendation tasks.
- The model effectively models matches between medical cases and syndrome elements.
- The integration of recalling and ranking stages improved recommendation accuracy.
Conclusions:
- The MMOE-match network offers a promising approach for enhancing recommendation systems in Traditional Chinese Medicine.
- This study highlights the potential of integrating advanced machine learning techniques like MTL and matching networks in TCM.
- Further research can explore manually designed matching features for potential improvements.
