Related Experiment Video
Updated: Jun 29, 2026
03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
Named entity recognition for coal mining machine assembly based on domain large language model.
Yunrui Wang1, Xintong Sui2, Zhaoyang Zheng2
1School of Mechanical Engineering, Xi'an University of Science and T echnology, Xi'an, 710054, China. wangyunr2001@xust.edu.cn.
Scientific Reports
|June 14, 2026
Summary
This study introduces a domain-specific large language model for Named Entity Recognition (NER) in coal mining machine assembly. The method significantly improves NER accuracy, aiding workers in understanding complex assembly data.
Area of Science:
- Engineering
- Computer Science
- Artificial Intelligence
Background:
- Coal mining machine assembly generates complex, unstructured text data, hindering operational efficiency.
- Traditional Named Entity Recognition (NER) models struggle with specialized, scarce datasets in this domain.
Purpose of the Study:
- To develop a domain-specific large language model (LLM) for improved NER in coal mining machine assembly.
- To enhance the understanding and processing of complex assembly-related textual data.
Main Methods:
- Comparative analysis of foundational LLMs to select the most suitable for coal mining domain data.
- Application of QLoRA (Quantized Low-Rank Adaptation) fine-tuning to optimize LLM parameters efficiently.
- Fine-tuning and evaluation on a real-world coal mining machine assembly dataset.
Main Results:
- Significant performance improvement in NER tasks, with BLEU-4 score increasing from 6.1225 to 65.8013.
- Achieved a high F1-Score of 0.893 for the NER task.
- Demonstrated effective handling of complex assembly data with reduced computational resource demands.
Conclusions:
- The proposed domain-specific LLM with QLoRA fine-tuning enhances NER accuracy for coal mining machine assembly.
- This advancement is crucial for assisting workers in comprehending intricate assembly information and boosting efficiency.