在自然語言處理中，解決問答問題是一項極具困難與挑戰的任務，這類型的任務需要進行語意理解，而後進行問題處理並進行正確的解釋輸出。尤其在進行數學文字題上的語意理解時，更需要具備數學知識，才能使語意表達的更為清楚，以便後續進行問題處理及解釋輸出時能更加精確。
本研究建立了一套系統進而分析小學數學文字題中的語意結構，藉由Pattern、知識圖譜來將文字題進行簡化，進而進行語意理解，並且得到語意結構。
作法主要包含：標註、簡化、產生語意結構。得到的語意結構，可以使解題器能夠直接取得所需要的重要資訊及對應的屬性，讓題目從計算機無法理解的句子中，轉變為可以進行分析的語意結構。而因使用簡化的原因，未來語意人員若要加入新的語意結構對應的Pattern就不需過度增加與考慮，只需考量簡化句的Pattern並加進系統後，即可使該系統可以處理新的問題類別。
本論文從2,327題的小學數學題目中，選出4,982個句子進行語意結構分析。將分析結果與經由哈工大剖析器與人工介入的標註，進行主詞、動詞、受詞的比對後，句子的正確率達87.6%。並且藉由該結果進行錯誤分析，發現可藉由其餘模組的效能改善，進而讓整體效能增進。

In natural language processing, question answering is a challenging task. Such task requires semantic understanding, then problem processing and getting correct interpretation output. Especially when understanding the semantics of mathematical word problems, it is even more necessary to have mathematical knowledge in order to express the semantics more clearly, so that the subsequent problem processing and interpretation output can be more accurate.
This study establishes a system to analyze the semantic structure of primary school mathematics word problems. It uses pattern and knowledge graph to simplify the problem sentences, and then understands semantic and gets the semantic structure.
The methods mainly include: labeling, reduction-based approach, generating semantic structure. It can transform the question from a sentence that cannot be understood by the computer into a semantic structure that can be analyzed.
Because of the reduction-based approach, if semantic researchers want to add patterns corresponding to new semantic structures in the future, they do not need to add too much. They only need to consider simplified sentence patterns and add them to the system, so that the system can handle new problem categories.
This paper selects 4,982 sentences from 2,327 elementary school mathematics word problem for semantic structure analysis. After comparing the analysis results with the labeling through the LTP parser and manual labeling. After we comparing the subject, verb, and object, the accuracy of the sentence reached 87.6%. And through error analysis based on the results, it was found that the system performance can be improved by improving the performance of other modules.

[1] S.Y. Miao, S.C. Lin, W.Y. Ma, and K.Y. Su. 構建一個中文國小數學文字問題語料庫(building a corpus for developing the Chinese elementary school math word problem solver)[in Chinese]. In Proceedings of the 28th Conference on Computational Linguistics and Speech Processing (ROCLING 2016), pages 352– 371, Tainan, Taiwan, Oct. 2016. The Association for Computational Linguistics and Chinese Language Processing (ACLCLP).
[2] A. Mukherjee and U. Garain. A review of methods for automatic understanding of natural language mathematical problems. Artificial Intelligence Review, 29(2):93– 122, Apr 2008.
[3] D. Zhang, L. Wang, L. Zhang, B. T. Dai, and H. T. Shen. The gap of semantic parsing: A survey on automatic math word problem solvers, 2019.
[4] S. Mandal and S. K. Naskar. Classifying and solving arithmetic math word problems— an intelligent math solver. IEEE Transactions on Learning Technologies, 14(1):28–41, 2021.
[5] C. Manning, M. Surdeanu, J. Bauer, J. Finkel, S. Bethard, and D. McClosky. The Stanford CoreNLP natural language processing toolkit. In Proceedings of 52nd Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 55–60, Baltimore, Maryland, June 2014. Association for Computational Linguistics.
[6] P. Qi, Y. Zhang, Y. Zhang, J. Bolton, and C. D. Manning. Stanza: A Python natural language processing toolkit for many human languages. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations, 2020. Association for Computational Linguistics.
[7] W.Y. Ma and K.J. Chen. Introduction to CKIP Chinese word segmentation system for the first international Chinese word segmentation bakeoff. In Proceedings of the Second SIGHAN Workshop on Chinese Language Processing, pages 168–171, Sapporo, Japan, July 2003. Association for Computational Linguistics.
[8] P.H. Li, T.J. Fu, and W.Y. Ma. Why Attention? Analyze BiLSTM Deficiency and Its Remedies in the Case of NER. In Proceedings of the Thirty-Third AAAI Conference on Artificial Intelligence.
[9] CKIP. CKIP Transformers. https://github.com/ckiplab/ckip-transformers, 2020.
[10] 詞庫小組. 詞庫小組技術報告9305 中文詞類分析(三版). 中央研究院資訊科 學研究所中文詞知識庫小組, 台北南港, 3 edition, June 1993.
[11] Y. Zhang and Z. Sui, editors. Proceedings of the 9th SIGHAN Workshop on Chinese Language Processing, Taiwan, Dec. 2017. Association for Computational Linguistics.
[12] G. Lample, M. Ballesteros, S. Subramanian, K. Kawakami, and C. Dyer. Neural architectures for named entity recognition, arXiv preprint arXiv:1603.01360, 2016.
[13] D. Marcheggiani and I. Titov, Encoding sentences with graph convolutional networks for semantic role labeling, arXiv preprint arXiv:1703.04826, 2017.
[14] F. Alva-Manchego, C. Scarton, and L. Specia, Data-Driven Sentence Simplification: Survey and Benchmark, Computational Linguistics, vol. 46, no. 1, pp. 135–187, 2020.
[15] T. J. Osborne, Tests for constituents: What they really reveal about the nature of syntactic structure, Language under discussion, vol. 5, no. 1, pp. 1–41, 2018.
[16] I. A. Mel’cuk et al., Dependency syntax: theory and practice. SUNY press, 1988.
[17] I. Sakai, Syntax in universal translation. Her Magesty’s Stationary Office, 1962.
[18] O. Etzioni, M. Banko, S. Soderland, and D. S. Weld, Open information extraction from the web, Communications of the ACM, vol. 51, no. 12, pp. 68–74, 2008.
[19] F. Wu and D. S. Weld, Open information extraction using wikipedia, in Proceedings of the 48th annual meeting of the association for computational linguistics, pp. 118–127, 2010. Association for Computational Linguistics.
[20] M. Schmitz, S. Soderland, R. Bart, O. Etzioni, et al., Open language learning for information extraction, in Proceedings of the 2012 joint conference on empirical methods in natural language processing and computational natural language learning, pp. 523–534, 2012. Association for Computational Linguistics.
[21] T. Gruber. Ontology. Encyclopedia of database systems, 1:1963–1965, 2009.
[22] T. R. Gruber. Toward principles for the design of ontologies used for knowledge sharing? International Journal of Human¬-Computer Studies, 43(5):907–928, 1995.
[23] W.L. Hsu, S.H. Wu, and Y.S. Chen. Event identification based on the information map-INFOMAP. In 2001 IEEE International Conference on Systems, Man and Cybernetics. e-Systems and e-Man for Cybernetics in Cyberspace (Cat.No.01CH37236), volume 3, pages 1661–1666 vol.3, 2001.
[24] H.¬J. Dai, C.¬K. Wang, N.¬W. Chang, M.¬S. Huang, J. Jonnagaddala, F.¬D. Wang, and W.¬L. Hsu. Statistical principle-¬based approach for recognizing and normalizing microRNAs described in scientific literature. Database, 2019, 2019.
[25] T.¬H. Yang, Y.¬L. Hsieh, S.¬H. Liu, Y.¬C. Chang, and W.¬L. Hsu. A flexible template generation and matching method with applications for publication reference meta¬ data extraction. Journal of the Association for Information Science and Technology, 72(1):32–45, 2021.