基于多维度特征融合及大语言模型技术的工业品物料分类系统

肖成祥

doi:10.3969/j.issn.2096-6091.2023.17.006

您当前的位置：

首页 >

文章列表页 >

基于多维度特征融合及大语言模型技术的工业品物料分类系统

科研通信 | 更新时间：2024-01-08

- 基于多维度特征融合及大语言模型技术的工业品物料分类系统
- A Material Classification System for Industrial Products Based on Multidimensional Feature Fusion Method and Large Language Model Technology
- 新一代信息技术 2023年6卷第17期页码：39-44
- 作者机构：
  
  欧冶工业品股份有限公司，上海 201900
- 作者简介：
  
  [ "肖成祥（1970—），男，出生于上海，本科学历，研究方向：人工智能、工业互联网、绿色低碳。" ]
- 基金信息：
- DOI：10.3969/j.issn.2096-6091.2023.17.006
  中图分类号：
扫描看全文
肖成祥. 基于多维度特征融合及大语言模型技术的工业品物料分类系统[J]. 新一代信息技术, 2023, 6(17): 39-44

XIAO Cheng-xiang. A Material Classification System for Industrial Products Based on Multidimensional Feature Fusion Method and Large Language Model Technology[J]. New Generation of Information Technology, 2023, 6(17): 39-44
肖成祥. 基于多维度特征融合及大语言模型技术的工业品物料分类系统[J]. 新一代信息技术, 2023, 6(17): 39-44 DOI： 10.3969/j.issn.2096-6091.2023.17.006.

XIAO Cheng-xiang. A Material Classification System for Industrial Products Based on Multidimensional Feature Fusion Method and Large Language Model Technology[J]. New Generation of Information Technology, 2023, 6(17): 39-44 DOI： 10.3969/j.issn.2096-6091.2023.17.006.

摘要

随着工业互联网的快速发展，工业品物料分类任务在精准供需匹配和数据规范化方面都具有重要意义。然而，传统的物料分类方法高度依赖人工经验与传统技术手法，效率和准确度有限。本研究提出了一种基于多维度特征融合及大语言模型技术的工业品物料分类系统，将深度学习与领域预训练相结合，提升了领域适应性和语义理解能力。采用了多维度特征融合手法、检索系统召回验证机制及大语言模型关键实体抽取技术，形成了一种更加精确、稳定、高效的工业品物料分类解决方案，助力了工业品供应链采购寻源等业务的数字化转型。

Abstract

With the rapid development of the Industrial Internet, the categorization of materials in industrial supply chain plays a crucial role in precise supply-demand matching and data standardization. However, traditional material classification methods heavily rely on manual experience and conventional techniques, resulting in limited efficiency and accuracy. This study proposes a material classification system based on multi-dimensional feature fusion and state-of-the-art large language model technology. By integrating deep learning with domain pre-training, the system enhances domain adaptability and semantic understanding. The approach employs multi-dimensional feature fusion, a retrieval system recall verification mechanism, and large language model key entity extraction technology. This combination forms a more precise, stable, and efficient solution for industrial material classification, thereby facilitating the digital transformation of industrial supply chain operations such as procurement and sourcing. This research contributes to the advancement of digitalization in the industrial sector.

关键词

数字化转型工业品供应链物料分类特征融合大语言模型

Keywords

digital transformationindustrial supply chainmaterial classificationfeature fusionlarge language model

references

DEVLIN J, CHANG M W, LEE K, et al. BERT: Pre-training of deep bidirectional transformers for language understanding[J]. Journal of Computer Science and Technology, 2018, 41(10): 1933-1947.

TENNEY I, DAS D, PAVLICK E. BERT rediscovers the classical NLP pipeline[EB/OL]. (2019-05-15)[2023-09-01]. https://arxiv.org/abs/1905.05950https://arxiv.org/abs/1905.05950.

SU P, VIJAY-SHANKER K. Investigation of improving the pre-training and fine-tuning of BERT model for biomedical relation extraction[J]. BMC Bioinformatics, 2022, 23(1): 120.

VASWANI A, et al. Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017, 30(8): 5998-6008.

KIM Y. Convolutional neural networks for sentence classification[EB/OL]. (2014-08-31)[2023-09-01]. https://arxiv.org/abs/1408.5882https://arxiv.org/abs/1408.5882.

GUO C, BERKHAHN F. Entity embeddings of categorical variables[EB/OL]. (2016-04-21)[2023-09-01]. https://arxiv.org/abs/1604.06737https://arxiv.org/abs/1604.06737.

HONGYANG L I, JUN C, RUIMIN H U. Multiple feature fusion in convolutional neural networks for action recognition[J]. Wuhan University Journal of Natural Sciences, 2017, 22(01): 78-83.

LAMPERT C H, NICKISCH H, HARMELING S. Attribute-based classification for zero-shot visual object categorization[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2014, 36(3): 453-465.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

暂无数据