浏览全部资源
扫码关注微信
1.广东工业大学 计算机学院,广东 广州 510006
2.佛山职业技术学院 人工智能学院,广东 佛山 528137
3.广东工业大学 管理学院,广东 广州 510006
4.广东省科技创新监测研究中心,广东 广州 510033
邓家亮(1995-),男,广东工业大学计算机学院硕士研究生,佛山职业技术学院助教,研究方向为自然语言处理、深度学习
祝家东(1989-),男,硕士,佛山职业技术学院副教授,研究方向为深度学习、大模型
邵长城(1992-),男,广东工业大学管理学院博士研究生,研究方向为推荐系统、深度学习、大模型
陈国梁(1982-),男,广东省科技创新监测研究中心工程师,研究方向为推荐系统、政务信息化
陈平华(1967-),男,硕士,广东工业大学计算机学院教授、硕士生导师,研究方向为大数据、推荐系统、人工智能。本文
收稿:2025-10-09,
纸质出版:2026-04-15
移动端阅览
邓家亮,祝家东,邵长城等.基于预训练模型的专家匹配任务分类重构[J].软件导刊,2026,25(04):123-131.
DENG Jialiang,ZHU Jiadong,SHAO Changcheng,et al.Classification-based Reframing of Expert Matching Tasks Powered by Pretrained Models[J].Software Guide,2026,25(04):123-131.
邓家亮,祝家东,邵长城等.基于预训练模型的专家匹配任务分类重构[J].软件导刊,2026,25(04):123-131. DOI: 10.11907/rjdk.251554.
DENG Jialiang,ZHU Jiadong,SHAO Changcheng,et al.Classification-based Reframing of Expert Matching Tasks Powered by Pretrained Models[J].Software Guide,2026,25(04):123-131. DOI: 10.11907/rjdk.251554.
在要求“小同行”评审的科研项目中,传统的专家匹配方法通常依赖项目描述与专家简历之间的双文本语义对齐,但在实际应用中面临专家简历数据稀缺、跨学科术语语义漂移等问题。为此,提出一种基于预训练模型的匹配任务分类重构模型:基于论文文本—专家标签数据对,训练一个从论文文本到专家标签的神经网络,将项目—专家匹配任务转化为“文本到标签”的分类问题。模型仅需项目描述作为输入,直接输出专家身份标签,从而绕过简历依赖并降低术语对齐难度。为解决中文专家标注数据稀缺问题,引入情感分类作为代理任务,系统评估BERT-base-Chinese、CNN、LSTM、RNN与SVM/TF-IDF五类模型的性能差异。实验表明,BERT-base-Chinese在自建跨学科专家数据集上的准确率和F1_weighted值均达到97%以上,明显优于其他模型,其全词掩码技术与分层微调策略能有效解决术语语义漂移问题。
In research project evaluation scenarios requiring "small-circle peer review," traditional expert retrieval/matching methods rely on semantic alignment between project descriptions and expert resumes(dual-text), facing practical challenges such as scarcity of expert resume data and semantic drift of cross-disciplinary terms. To address these issues, this paper proposes a classification-based reframing model for matching tasks using pretrained models. Based on paper text-expert label data pairs, the model trains a neural network that maps paper text to expert labels, transforming the project-expert matching task into a "text-to-label" classification problem. The model requires only the project description as input and directly outputs expert identity labels, thereby bypassing the dependency on resumes and reducing the difficulty of term alignment. To mitigate the scarcity of annotated Chinese expert data, sentiment classification is introduced as a proxy task to systematically evaluate the performance differences among five models: BERT-base-Chinese, CNN, LSTM, RNN, and SVM/TF-IDF. Experiments show that BERT-base-Chinese achieves an accuracy rate above 97% and an F1_weighted-score above 97% on a self-constructed cross-disciplinary dataset, significantly outperforming other models. Its Whole Word Masking(WWM) technique and layered fine-tuning strategy effectively resolve term semantic drift.
JIANG X Y , LI D L . Research on the construction of a multi-level intelligent project review expert management system: optimization based on the management of scientific research project review experts in Guangdong [J]. Science and Technology Management Research , 2018 , 38 ( 16 ): 212 - 217 . doi: 10.3969/j.issn.1000-7695.2018.16.031 http://dx.doi.org/10.3969/j.issn.1000-7695.2018.16.031
江笑颜 , 李栋亮 . 多层次、智能化的项目评审专家管理体系构建研究——基于广东科研项目评审专家管理的优化 [J]. 科技管理研究 , 2018 , 38 ( 16 ): 212 - 217 . doi: 10.3969/j.issn.1000-7695.2018.16.031 http://dx.doi.org/10.3969/j.issn.1000-7695.2018.16.031
YU Y , FU Y , WU X P . A survey of Chinese text classification methods [J]. Journal of Cyber Security , 2019 , 5 ( 5 ): 1 - 8 . doi: 10.11959/j.issn.2096-109x.2019045 http://dx.doi.org/10.11959/j.issn.2096-109x.2019045
于游 , 付钰 , 吴晓平 . 中文文本分类方法综述 [J]. 网络与信息安全学报 , 2019 , 5 ( 5 ): 1 - 8 . doi: 10.11959/j.issn.2096-109x.2019045 http://dx.doi.org/10.11959/j.issn.2096-109x.2019045
KUNCORO B A , ISWANTO B H . TF-IDF method in ranking keywords of Instagram users' image captions [C]// Piscataway:Proceedings of 2015 International Conference on Information Technology Systems and Innovation (ICITSI) , 2016 . doi: 10.1109/icitsi.2015.7437705 http://dx.doi.org/10.1109/icitsi.2015.7437705
ZHENG Y , MENG Z P , XU C . A short-text oriented clustering method for hot topics extraction [J]. International Journal of Software Engineering and Knowledge Engineering , 2015 , 25 ( 3 ): 453 - 471 . doi: 10.1142/s0218194015400161 http://dx.doi.org/10.1142/s0218194015400161
LIN Y , ZHANG J S , LIU Z H , et al . Application of technical review expert selection method in disruptive technology expert prediction platform [J]. China Science & Technology Resources Review , 2024 , 56 ( 2 ): 54 - 62 .
林毅 , 张均胜 , 刘志辉 , 等 . 技术评审专家遴选方法在颠覆性技术专家预判平台上的应用 [J]. 中国科技资源导刊 , 2024 , 56 ( 2 ): 54 - 62 .
XIE X J , LIANG Y , WANG Z S , et al . Semantic matching method for academic expertise of experts integrating multi-source information [J]. Chinese Journal of Computers , 2023 , 46 ( 5 ): 1045 - 1059 .
谢小杰 , 梁英 , 王梓森 , 等 . 融合多源信息的专家学术专长语义匹配方法 [J]. 计算机学报 , 2023 , 46 ( 5 ): 1045 - 1059 .
WEI X L , SUN X . Research progress and development trends of natural language processing technologies and methods: a review [J]. China-Arab Science and Technology Forum (Chinese and English) , 2025 ( 5 ): 84 - 88 .
卫宣伶 , 孙翔 . 自然语言处理技术方法研究进展与发展趋势综述 [J]. 中阿科技论坛(中英文) , 2025 ( 5 ): 84 - 88 .
MIKOLOV T , SUTSKEVER I , CHEN K , et al . Distributed representations of words and phrases and their compositionality[DB/OL] . https://arxiv.org/abs/1310.4546 https://arxiv.org/abs/1310.4546 . doi: 10.32614/cran.package.word2vec http://dx.doi.org/10.32614/cran.package.word2vec
CHANG J Q , SHEN W . Research on Chinese word segmentation algorithm based on string matching [J]. Industrial Control Computer , 2016 , 29 ( 2 ): 115 - 116,119 . doi: 10.3969/j.issn.1001-182X.2016.02.051 http://dx.doi.org/10.3969/j.issn.1001-182X.2016.02.051
常建秋 , 沈炜 . 基于字符串匹配的中文分词算法的研究 [J]. 工业控制计算机 , 2016 , 29 ( 2 ): 115 - 116,119 . doi: 10.3969/j.issn.1001-182X.2016.02.051 http://dx.doi.org/10.3969/j.issn.1001-182X.2016.02.051
MALDONADO S , LÓPEZ J . Dealing with high-dimensional class-imbalanced datasets: embedded feature selection for SVM classification [J]. Applied Soft Computing , 2018 , 67 : 94 - 105 . doi: 10.1016/j.asoc.2018.02.051 http://dx.doi.org/10.1016/j.asoc.2018.02.051
SALTON G , WONG A , YANG C S . A vector space model for automatic indexing [J]. Communications of the ACM , 1975 , 18 ( 11 ): 613 - 620 . doi: 10.1145/361219.361220 http://dx.doi.org/10.1145/361219.361220
JACCARD P . The distribution of the flora in the alpine zone [J]. New Phytologist , 1912 , 11 ( 2 ): 37 - 50 . doi: 10.2307/2427226 http://dx.doi.org/10.2307/2427226
HE Y . Research on support vector machines in text classification [J]. Henan Science and Technology , 2019 ( 29 ): 8 - 10 .
何焱 . 文本分类中支持向量机研究 [J]. 河南科技 , 2019 ( 29 ): 8 - 10 .
YANG Q , WANG A Z , WU J T , et al . A survey of dichotomous image segmentation based on deep learning [J/OL]. Computer Engineering and Applications , 2025-09-14 . https://link.cnki.net/urlid/11.2127.TP.20250818.1649.004 https://link.cnki.net/urlid/11.2127.TP.20250818.1649.004 . doi: 10.3778/j.issn.1002-8331.2504-0200 http://dx.doi.org/10.3778/j.issn.1002-8331.2504-0200
杨茜 , 王安志 , 吴锦涛 , 等 . 基于深度学习的二分图像分割综述 [J/OL]. 计算机工程与应用 , 2025-09-14 . https://link.cnki.net/urlid/11.2127.TP.20250818.1649.004 https://link.cnki.net/urlid/11.2127.TP.20250818.1649.004 . doi: 10.3778/j.issn.1002-8331.2504-0200 http://dx.doi.org/10.3778/j.issn.1002-8331.2504-0200
KANG X X , NIE Y . The construction and application of dual-teacher teaching model for enhancing students' deep learning capabilities:from perspective of ICAP theory [J]. China Educational Technology , 2025 ( 9 ): 126 - 135 .
康欣欣 , 乜勇 . 面向学生深度学习能力提升的双师教学模式构建及应用研究——基于ICAP理论的视角 [J]. 中国电化教育 , 2025 ( 9 ): 126 - 135 .
GONG L N , ZHOU Y R , QIAO Y , et al . Research progress on the application of pre-trained models in the field of software engineering [J]. Journal of Software , 2025 , 36 ( 1 ): 1 - 26 .
宫丽娜 , 周易人 , 乔羽 , 等 . 预训练模型在软件工程领域应用研究进展 [J]. 软件学报 , 2025 , 36 ( 1 ): 1 - 26 .
VASWANI A , SHAZEER N , PARMAR N , et al . Attention is all you need [C]// Proceeding of the 31st International Conference on Neural Information Processing Systems , 2017 : 1 - 11 .
DEVLIN J , CHANG M W , LEE K , et al . Bert: pre-training of deep bidirectional transformers for language understanding[DB/OL] . https://arxiv.org/abs/1810.04805 https://arxiv.org/abs/1810.04805 .
HU B , LU Z , LI H , et al . Convolutional neural network architectures for matching natural language sentences[DB/OL] . https://arxiv.org/abs/1503.03244 https://arxiv.org/abs/1503.03244 . doi: 10.7717/peerj-cs.616/fig-2 http://dx.doi.org/10.7717/peerj-cs.616/fig-2
SEPP H , JÜRGEN S . Long short-term memory [J]. Neural Computation , 1997 , 9 ( 8 ): 1735 - 1780 . doi: 10.1162/neco.1997.9.8.1735 http://dx.doi.org/10.1162/neco.1997.9.8.1735
PAN S J , YANG Q . A survey on transfer learning [J]. IEEE Transactions on Knowledge and Data Engineering , 2010 , 22 ( 10 ): 1345 - 1359 . doi: 10.1109/tkde.2009.191 http://dx.doi.org/10.1109/tkde.2009.191
SUN C , QIU X , XU Y , et al . How to fine-tune bert for text classification? [C]// China National Conference on Chinese Computational Linguistics , 2019 : 194 - 206 . doi: 10.1007/978-3-030-32381-3_16 http://dx.doi.org/10.1007/978-3-030-32381-3_16
KIM Y . Convolutional neural networks for sentence classification [C]// Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) , 2014 : 1746 - 1751 . doi: 10.3115/v1/d14-1181 http://dx.doi.org/10.3115/v1/d14-1181
0
浏览量
1
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
鄂公网安备42010602004949号