Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Knowledge-Enhanced Disease Diagnosis Method Based on Prompt Learning and BERT Integration

Zhang Zheng

TL;DR

This work addresses the challenge of structured reasoning in disease diagnosis by marrying external knowledge graphs with a prompt-learning framework and BERT. It retrieves and encodes structured clinical knowledge, injects it into prompt templates, and uses a verbalizer to map model outputs to diagnoses, achieving state-of-the-art results on CHIP-CTC, IMCS-V2-NER, and KUAKE-QTR with $F1$ gains of $2.4\%$, $3.1\%$, and $4.2\%$, respectively. Ablation studies confirm the essential role of the knowledge-injection component and highlight the sensitivity to prompt design and verbalizers. The approach also enhances interpretability by exposing the knowledge paths that underlie each diagnosis, supporting clinical decision-making with traceable evidence.

Abstract

This paper proposes a knowledge-enhanced disease diagnosis method based on a prompt learning framework. The method retrieves structured knowledge from external knowledge graphs related to clinical cases, encodes it, and injects it into the prompt templates to enhance the language model's understanding and reasoning capabilities for the task.We conducted experiments on three public datasets: CHIP-CTC, IMCS-V2-NER, and KUAKE-QTR. The results show that the proposed method significantly outperforms existing models across multiple evaluation metrics, with an F1 score improvement of 2.4% on the CHIP-CTC dataset, 3.1% on the IMCS-V2-NER dataset,and 4.2% on the KUAKE-QTR dataset. Additionally,ablation studies confirmed the critical role of the knowledge injection module,as the removal of this module resulted in a significant drop in F1 score. The experimental results demonstrate that the proposed method not only effectively improves the accuracy of disease diagnosis but also enhances the interpretability of the predictions, providing more reliable support and evidence for clinical diagnosis.

A Knowledge-Enhanced Disease Diagnosis Method Based on Prompt Learning and BERT Integration

TL;DR

This work addresses the challenge of structured reasoning in disease diagnosis by marrying external knowledge graphs with a prompt-learning framework and BERT. It retrieves and encodes structured clinical knowledge, injects it into prompt templates, and uses a verbalizer to map model outputs to diagnoses, achieving state-of-the-art results on CHIP-CTC, IMCS-V2-NER, and KUAKE-QTR with gains of , , and , respectively. Ablation studies confirm the essential role of the knowledge-injection component and highlight the sensitivity to prompt design and verbalizers. The approach also enhances interpretability by exposing the knowledge paths that underlie each diagnosis, supporting clinical decision-making with traceable evidence.

Abstract

This paper proposes a knowledge-enhanced disease diagnosis method based on a prompt learning framework. The method retrieves structured knowledge from external knowledge graphs related to clinical cases, encodes it, and injects it into the prompt templates to enhance the language model's understanding and reasoning capabilities for the task.We conducted experiments on three public datasets: CHIP-CTC, IMCS-V2-NER, and KUAKE-QTR. The results show that the proposed method significantly outperforms existing models across multiple evaluation metrics, with an F1 score improvement of 2.4% on the CHIP-CTC dataset, 3.1% on the IMCS-V2-NER dataset,and 4.2% on the KUAKE-QTR dataset. Additionally,ablation studies confirmed the critical role of the knowledge injection module,as the removal of this module resulted in a significant drop in F1 score. The experimental results demonstrate that the proposed method not only effectively improves the accuracy of disease diagnosis but also enhances the interpretability of the predictions, providing more reliable support and evidence for clinical diagnosis.
Paper Structure (18 sections, 12 equations, 5 figures, 3 tables)

This paper contains 18 sections, 12 equations, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. Disease Diagnosis
  4. Application of the BERT Model in Disease Diagnosis
  5. Task formulation
  6. Methodology
  7. Knowledge Retrieval
  8. Knowledge Representation
  9. Prompt Template
  10. Prediction
  11. Experimental settings
  12. Datasets
  13. Baseline
  14. Results and Analysis
  15. Comparison Experiments
  16. ...and 3 more sections

Figures (5)

  • Figure 1: Method Structure Diagram
  • Figure 2: Performance of Different Benchmark Models and the Proposed Model on the KUAKE-QTR Dataset
  • Figure 3: Performance of Different Benchmark Models and the Proposed Model on the IMCS-V2-NER Dataset
  • Figure 4: Performance of Different Benchmark Models and the Proposed Model on the CHIP-CTC Dataset
  • Figure 5: Accuracy Comparison of Eight Models Across Three Datasets