基于知识库问答的回答生成研究

doi:10.19678/j.issn.1000-3428.0068433

摘要/Abstract

摘要：

知识库问答旨在利用事先构建好的知识库来回答用户提出的问题。现有的知识库问答研究主要通过对候选实体和关系路径进行排序, 最后将三元组的尾实体作为答案返回。用户给出的问题经过实体识别模型和实体消歧模型之后, 可以链接到知识库中与答案相关的候选实体。利用语言模型的生成能力, 可以将答案拓展为一句话并返回, 这对用户而言是更加友好的。为了提高模型的泛化能力和弥补问题文本与结构化知识之间的差别, 将候选实体及其一跳关系子图通过提示模板进行组织输入到生成模型中, 并在回答模板的引导下生成通俗流畅的回答。在NLPCC 2016 CKBQA和KgCLUE两个中文数据集上的实验结果表明: 该方法在BLEU、METEOR和ROUGE指标上分别平均比BART-large模型提高了2.8、2.3和1.5百分点; 在Perplexity指标上, 该方法与ChatGPT的回答表现相当。

关键词: 知识库问答, 提示, 实体链接, 预训练模型, 回答生成

Abstract:

Knowledge base question answering aims to use pre-constructed knowledge bases to answer questions raised by users. Existing knowledge base question answering research sorts candidate entities and relationship paths and finally returns the tail entity of the triple as the answer. After the questions provided by the user pass through the entity recognition and entity disambiguation models, they can be linked to candidate entities related to the answers in the knowledge base. Using the generation capability of the language model, the answer can be expanded into a sentence and returned, which is more user-friendly. To improve the generalization ability of the model and compensate for the difference between the question text and structured knowledge, candidate entities and their one-hop relationship subgraphs are organized and input into the generation model through a prompt template, and a popular and fluent text is generated under the guidance of the answer template. Experimental results on the NLPCC 2016 CKBQA and KgCLUE Chinese datasets indicated that on average, the proposed method outperforms the BART-large model by 2.8, 2.3, and 1.5 percentage points on the Bilingual Evaluation Understudy (BLEU), Metric for Evaluation of Translation with Explicit Ordering (METEOR), and Recall-Oriented Understudy for Gisting Evaluation (ROUGE) series metrics, respectively. For the Perplexity metric, the method performs comparably to the ChatGPT responses.

Key words: knowledge base question answering, prompt, entity linking, pre-training language model, answer generation

饶东宁, 许正辉, 梁瑞仕. 基于知识库问答的回答生成研究[J]. 计算机工程, 2025, 51(2): 94-101.

RAO Dongning, XU Zhenghui, LIANG Ruishi. Research on Answer Generation Based on Knowledge Base Question Answering[J]. Computer Engineering, 2025, 51(2): 94-101.

收藏文章 0 / 推荐 / 导出引用

链接本文: https://www.ecice06.com/CN/10.19678/j.issn.1000-3428.0068433

https://www.ecice06.com/CN/Y2025/V51/I2/94

图/表 13

图1 模型结构

Fig.1 Model structure

图2 实体识别模型

Fig.2 Entity recognition model

图3 实体消歧模型

Fig.3 Entity disambiguation model

图4 生成模型

Fig.4 Generation model

参考文献 35

1	DENG C Y , ZENG G F , CAI Z P , et al. A survey of knowledge based question answering with deep learning. Journal on Artificial Intelligence, 2020, 2 (4): 157- 166. doi: 10.32604/jai.2020.011541
2	HUANG X, KIM J J, ZOU B W. Unseen entity handling in complex question answering over knowledge base via language generation[C]//Proceedings of EMNLP 2021. Stroudsburg, USA: Association for Computational Linguistics, 2021: 547-557. 10.18653/v1/2021.findings-emnlp.50
3	PÉREZ J , ARENAS M , GUTIERREZ C . Semantics and complexity of SPARQL. ACM Transactions on Database Systems, 2009, 34 (3): 1- 45. URL
4	YE X, YAVUZ S, HASHIMOTO K, et al. RNG-KBQA: generation augmented iterative ranking for knowledge base question answering[C]//Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics(Volume 1: Long Papers). Stroudsburg, USA: Association for Computational Linguistics, 2022: 6032-6043.10.18653/v1/2022.acl-long.417
5	HU X, WU X, SHU Y, et al. Logical form generation via multi-task learning for complex question answering over knowledge bases[C]//Proceedings of the 29th International Conference on Computational Linguistics. Gyeongju, Republic of Korea: [s. n. ], 2022: 1687-1696.
6	何展鹏. 基于知识库标记预训练孪生神经网络的中文实体链接. 计算机科学与应用, 2022, 12 (4): 1202- 1212. doi: 10.12677/CSA.2022.124122
	HE Z P . Knowledge marker-based pre-trained language model with siamese network for Chinese entity linking. Computer Science and Application, 2022, 12 (4): 1202- 1212. doi: 10.12677/CSA.2022.124122
7	XUE L T, CONSTANT N, ROBERTS A, et al. mT5: a massively multilingual pre-trained text-to-text transformer[EB/OL]. [2023-05-10]. https://arxiv.org/pdf/2010.11934.
8	ROSSIELLO G, MIHINDUKULASOORIYA N, ABDELAZIZ I, et al. Generative relation linking for question answering over knowledge bases[C]//Proceedings of International Semantic Web Conference. Berlin, Germany: Springer, 2021: 321-337. 10.48550/arXiv.2108.07337
9	DUAN N. Overview of the NLPCC-ICCPOL 2016 shared task: open domain Chinese question answering[C]//Proceedings of NLPCC 2016 and ICCPOL 2016. Berlin, Germany: Springer, 2016: 942-948. 10.1007/978-3-319-50496-4_89
10	DONG G T, LI R M, WANG S R, et al. Bridging the KB-text gap: leveraging structured knowledge-aware pre-training for KBQA[C]//Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. New York, USA: ACM Press, 2023: 3854-3859.
11	COLIN R , NOAM S , ADAM R , et al. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of Machine Learning Research, 2020, 21 (1): 5485- 5551.
12	LEWIS M, LIU Y H, GOYAL N, et al. BART: denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg, USA: Association for Computational Linguistics, 2020: 7871-7880.
13	ZHANG L X, ZHANG J, WANG Y L, et al. FC-KBQA: a fine-to-coarse composition framework for knowledge base question answering[EB/OL]. [2023-05-10]. https://arxiv.org/pdf/2306.14722.
14	SEVGILI Ö , SHELMANOV A , ARKHIPOV M , et al. Neural entity linking: a survey of models based on deep learning. Semantic Web, 2022, 13 (3): 527- 570. doi: 10.3233/SW-222986
15	CHEN Y , WAN W B , ZHAO Y M , et al. Generalization performance optimization of KBQA system for Chinese open domain. Multimedia Tools and Applications, 2024, 83 (5): 12445- 12466. URL
16	ZHANG W T, JIANG S H, ZHAO S, et al. A BERT-BiLSTM-CRF model for Chinese electronic medical records named entity recognition[C]//Proceedings of the 12th International Conference on Intelligent Computation Technology and Automation (ICICTA). Washington D. C., USA: IEEE Press, 2019: 166-169. 10.1109/ICICTA49267.2019.00043
17	KANNAN RAVI M P, SINGH K, MULANG' I O, et al. CHOLAN: a modular approach for neural entity linking on Wikipedia and Wikidata[C]//Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume. Stroudsburg, USA: Association for Computational Linguistics, 2021: 504-514. 10.48550/arXiv.2101.09969
18	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[EB/OL]. [2023-05-10]. https://arxiv.org/abs/1706.03762.
19	LIU P F , YUAN W Z , FU J L , et al. Pre-train, prompt, and predict: a systematic survey of prompting methods in natural language processing. ACM Computing Surveys, 2023, 55 (9): 1- 35. URL
20	SCHICK T, SCHVTZE H. Exploiting cloze-questions for few-shot text classification and natural language inference[EB/OL]. [2023-05-10]. https://arxiv.org/abs/2001.07676.
21	LI X L, LIANG P. Prefix-tuning: Optimizing continuous prompts for generation[EB/OL]. [2023-05-10]. https://arxiv.org/abs/2101.00190.
22	GAO T Y, FISCH A, CHEN D Q. Making pre-trained language models better few-shot learners[EB/OL]. [2023-05-10]. https://arxiv.org/abs/2012.15723.
23	CHEN Y L, LIU Y, DONG L, et al. AdaPrompt: adaptive model training for prompt-based NLP[EB/OL]. [2023-05-10]. https://arxiv.org/abs/2202.04824.
24	ZHONG W J, GAO Y F, DING N, et al. ProQA: structural prompt-based pre-training for unified question answering[EB/OL]. [2023-05-10]. https://arxiv.org/abs/2205.04040.
25	LV X, LIN Y K, CAO Y X, et al. Do pre-trained models benefit knowledge graph completion? A reliable evaluation and a reasonable approach[C]//Proceedings of the Findings of the Association for Computational Linguistics. Stroudsburg, USA: Association for Computational Linguistics, 2022: 1-10. 10.18653/v1/2022.findings-acl.282
26	TAN C Y, CHEN Y H, SHAO W B, et al. Make a choice! Knowledge base question answering with In-context learning[EB/OL]. [2023-05-10]. https://arxiv.org/abs/2305.13972.
27	DEVLIN J, CHANG M W, LEE K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[EB/OL]. [2023-05-10]. https://arxiv.org/abs/1810.04805.
28	TJONG KIM SANG E F, de MEULDER F. Introduction to the CoNLL-2003 shared task: language-independent named entity recognition[C]//Proceedings of the 17th Conference on Natural Language learning at HLT-NAACL 2003. Stroudsburg, USA: Association for Computational Linguistics, 2003: 1-10. 10.48550/arXiv.cs/0306050
29	REIMERS N, GUREVYCH I. Sentence-BERT: sentence embeddings using siamese BERT-networks[EB/OL]. [2023-05-10]. https://arxiv.org/abs/1908.10084.
30	LIU A T, HUANG Z Q, LU H T, et al. BB-KBQA: BERT-based knowledge base question answering[C]//Proceedings of China National Conference on Chinese Computational Linguistics. Berlin, Germany: Springer, 2019: 81-92. 10.1007/978-3-030-32381-3_7
31	潘春光, 党金明, 杨智, 等. CCKS 2019 & 百度2019中文短文本的实体链指第一名方案[EB/OL]. [2023-05-10]. https://github.com/panchunguang/ccks_baidu_entity_link.
	PAN C G, DANG J M, YANG Z, et al. The first place solution for entity linking in Chinese short texts in CCK & Baidu 2019[EB/OL]. [2023-05-10]. https://github.com/panchunguang/ccks_baidu_entity_link. (in Chinese)
32	LIU W J , ZHOU P , ZHAO Z , et al. K-BERT: enabling language representation with knowledge graph. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34 (3): 2901- 2908. doi: 10.1609/aaai.v34i03.5681
33	PAPINENI K, ROUKOS S, WARD T, et al. BLEU: a method for automatic evaluation of machine translation[C]//Proceedings of the 40th Annual Meeting on Association for Computational Linguistic. Stroudsburg, USA: Association for Computational Linguistics, 2001: 311-318.
34	LIN C Y. ROUGE: a package for automatic evaluation of summaries[EB/OL]. [2023-05-10]. https://www.semanticscholar.org/paper/ROUGE%3A-A-Package-for-Automatic-Evaluation-of-Lin/60b05f32c32519a809f21642ef1eb3eaf3848008?p2df.
35	BANERJEE S, LAVIE A. METEOR: an automatic metric for MT evaluation with improved correlation with human judgments[C]//Proceedings of the ACL Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and/or Summarization. Stroudsburg, USA: Association for Computational Linguistics, 2005: 65-72.

编辑推荐 0

Metrics

阅读次数

全文

HTML			PDF

最新录用	在线预览	正式出版	最新录用	在线预览	正式出版
0	0	3	23	0	52

来源	本网站	其他网站

次数	55	23
比例	71%	29%

摘要

160

最新录用	在线预览	正式出版

75	0	85

来源	本网站	其他网站

次数	37	123
比例	23%	77%

[1]	姚利峰, 蔡满春, 朱懿, 陈咏豪, 张溢文. 基于字节编码与预训练任务的加密流量分类模型[J]. 计算机工程, 2025, 51(2): 188-201.
[2]	费涛, 艾山·吾买尔, 杜文旭, 朱翠翠. 基于Squeezeformer的多颗粒度多方面发音质量评测方法[J]. 计算机工程, 2025, 51(1): 81-87.
[3]	魏嵬, 丁香香, 郭梦星, 杨钊, 刘辉. 文本相似度计算方法综述[J]. 计算机工程, 2024, 50(9): 18-32.
[4]	杨冬菊, 黄俊涛. 基于大语言模型的中文科技文献标注方法[J]. 计算机工程, 2024, 50(9): 113-120.
[5]	曾碧卿, 陈鹏飞, 姚勇涛. 融合思维链和低秩自适应微调的方面情感三元组抽取[J]. 计算机工程, 2024, 50(7): 53-62.
[6]	周炫余, 吴莲华, 郑勤华, 肖天星, 王紫璇, 张思敏. 联合语义提示和记忆增强的弱监督跳绳视频异常检测方法[J]. 计算机工程, 2024, 50(7): 87-95.
[7]	周昭辰, 方清茂, 吴晓红, 胡平, 何小海. 基于MacBERT与对抗训练的机器阅读理解模型[J]. 计算机工程, 2024, 50(5): 41-50.
[8]	李田芳, 普园媛, 赵征鹏, 徐丹, 钱文华. 基于CLIP和双空间自适应归一化的图像翻译[J]. 计算机工程, 2024, 50(5): 229-240.
[9]	邓远飞, 李加伟, 蒋运承. 基于知识注入提示学习的专利短语相似度计算[J]. 计算机工程, 2024, 50(4): 294-302.
[10]	侯钰涛, 阿布都克力木·阿布力孜, 史亚庆, 马依拉木·木斯得克, 哈里旦木·阿布都克里木. 面向"一带一路"的低资源语言机器翻译研究[J]. 计算机工程, 2024, 50(4): 332-341.
[11]	于明诚, 党亚固, 吴奇林, 吉旭, 毕可鑫. 基于多尺度上下文的英文作文自动评分研究[J]. 计算机工程, 2024, 50(3): 259-266.
[12]	陈志强, 仇瑜, 朱宇, 王晓英. 基于先验知识引导提示学习的自监督分类法补全[J]. 计算机工程, 2024, 50(12): 151-162.
[13]	张文博, 黄浩, 吴迪, 唐敏杰. 基于MEGA网络和分层预测的标点恢复方法[J]. 计算机工程, 2024, 50(12): 396-406.
[14]	孙仁科, 许靖昊, 皇甫志宇, 李仲年, 许新征. 基于视觉-语言预训练模型的零样本迁移学习方法综述[J]. 计算机工程, 2024, 50(10): 1-15.
[15]	曹发鑫, 孙媛媛, 王治政, 潘丁豪, 林鸿飞. 面向借贷案件的相似案例匹配模型[J]. 计算机工程, 2024, 50(1): 306-312.

选择文件类型/文献管理软件名称

选择包含的内容