進階搜尋


下載電子全文  
系統識別號 U0026-0308202009431000
論文名稱(中文) 使用具注意力機制之強化學習於商品之英文評論摘要生成方法-以Amazon電商平台為例
論文名稱(英文) A Reinforcement Learning Method with Attention Mechanism for Generating English Abstractive Summary of Products - Using Amazon E-Commerce as Examples
校院名稱 成功大學
系所名稱(中) 工程科學系
系所名稱(英) Department of Engineering Science
學年度 108
學期 2
出版年 109
研究生(中文) 詹定璿
研究生(英文) Ding-Xuan Zhan
電子信箱 afly.bsky@gmail.com
學號 N96071075
學位類別 碩士
語文別 中文
論文頁數 125頁
口試委員 指導教授-王宗一
共同指導教授-高宏宇
口試委員-賴槿峰
口試委員-李政德
口試委員-張家瑋
中文關鍵字 摘要生成  機器學習  強化學習  意見探勘  情感分析  數位媒體 
英文關鍵字 Automated Summary  Deep learning  Reinforcement Learning  Pointer Generator 
學科別分類
中文摘要 現代人為了方便,上網購物已經變成常態。消費者在購物平台上看到想購買的產品時,因為無法看到或實際試用該產品以做決定,通常會參考平台上該產品的顧客評論和摘要來做購買與否的決定。但通常平台上的顧客評論可能過於口語化,或是摘要過於簡略並沒有提到該產品的關鍵特徵及規格,導致消費者只知道產品很棒或是很差,但無法了解該產品的特徵是好還是壞,因此單單平台上的評論摘要通常無法滿足潛在顧客的要求。而本研究主要以亞馬遜網站的評論及摘要為資料,透過深度學習,針對不同類別的產品評論及摘要進行分析,並產生具關鍵特徵之摘要。本研究結合詞性標註、句法依賴及片語修飾關係找出評論中的關鍵詞,最後藉由機器學習文本關鍵詞與文本內容,從而理解產品評論中句子的語意,並生成簡單易懂的文本摘要,冀望能輔助消費者快速理解評論中的重要資訊。
本研究主要特點如下,1.針對評論文本設計文法及句法依賴規則,能針對不同類型評論語句提取關鍵詞,並可依據需求再進行規則擴充。2.修改原有的Attention機制改以加入Intra Attention機制之指標網路進行生成,使decoder在生成摘要詞彙時會重新考慮過去已生成序列所產生的temporal Attention scores,以避免模型在生成時過度關注相同的已生成詞彙。3.在原有Attention 的機制裡加入keyword的語意特徵,使計算出注意力權重比起原有的Attention機制更能集中在關鍵詞彙上。4.套用Self-Critical-Sequence-Training方法進一步優化Pointer-Base指標網路。本研究進行了十三個驗證,第一個驗證著重在關鍵字提取的準確性,第二到第四個驗證著重於分析模型加入不同Attention機制的詞彙分布,第五到第十一驗證則是與近年提出的抽取式和摘要式摘要方法比較準確性,第十二到第十三驗證則是以本論文最佳模型,在不同類型的產品評論進行評論摘要生成,並分別以Rouge、BLEU及METEOR三種方式進行準確性之評估。
英文摘要 Today shopping online has become a daily practice for many people. A customer, being interested in something on a web shop and not able to examine and try the real product, might turn to the custom reviews to see the opinions of buyers on the product before making the final decision. But the reviews may be too colloquial, or the summary may be too short to mention important key characteristics and the detail specs of the product. This makes the customer informed only that the product is good or bad, but not that important key characteristics of the product are good or bad. Such reviews of a web shop may not actually be helpful for its potential customers. This research mainly aims at Amazon review and summary information and uses deep learning approach to analyze and learn reviews and summaries of different classes of products and to produce summaries with import key characteristics of products for customers. The approach first uses part-of-speech (POS) tagging, syntactic dependence, and noun phrases to find the keywords in the reviews and summaries, and then uses reinforcement learning to learn the keywords and their review contents in order to understand the semantic of sentences in the customer reviews. Given a product review, the network produces a summary that is easy to read and contains information of important key characteristics of the product and is helpful for consumers to quickly understand the quality of the product and make the final decision.
The major research works of this study are as follows: 1. Design grammar and syntactic dependence rules of sentence and use them to extract keywords from different types of sentences in the reviews. Those rules can be expanded in the future when there are new requirements. 2. Replace the Attention mechanism of a Pointer-Generator with the Intra Attention mechanism to make the decoder reconsider the temporal Attention scores of previously generated sequences when generating summary vocabulary. This makes the model avoid focusing on a same word when the model is in generation mode. 3. Add extra keyword semantic information in the original Attention mechanism, which makes the generated attention weights focus more on the key words than the original Attention mechanism. 4. Apply the self-critical-sequence training method to optimize the Pointer-Generator network.
This study conducts totally thirteen experiments. The first experiment focuses on the accuracy of keyword extractions. The second to the fourth experiments focus on the analysis of the vocabulary distribution of different Attention mechanisms in the Pointer-Generator network. The fifth to the eleventh experiments are to compare the accuracy with some extractive and abstractive summary methods proposed in recent years. The twelfth to the thirteenth experiments, by using the best model in this study, generate summary for different types of product reviews and evaluate the accuracy of the model by three methods, namely Rouge, BLEU and METEOR.
論文目次 摘要 I
Extended Abstract II
致謝 VII
目錄 VIII
表目錄 XI
圖目錄 XIII
第一章 緒論 1
1.1 研究背景與動機 1
1.2 研究目的 6
1.3 研究方法 6
1.4 研究貢獻 7
第二章 文獻探討 8
2.1 評論關鍵詞提取 8
2.2 文本處理技術 10
2.3 語言特徵表示 11
2.4 自動摘要 13
2.5 注意力機制 16
2.6 指標網路結構 17
第三章 模型設計與架構 25
3.1 評論資料前處理 26
3.1.1 英文縮寫與還原 26
3.1.2 文本分詞 27
3.1.3 拼寫較正 30
3.1.4 詞性還原 32
3.1.5 情感分析及過濾 32
3.2 評論關鍵詞提取 34
3.3 資料流程整理 41
3.4 注意力機制方法 43
3.4.1 Keyword Attention 43
3.4.2 Intra-Temporal Attention 43
3.4.3 Sequential Intra-Attention 44
3.4.4 Self-Attention 46
3.4.5 Multi-Headed Attention 48
3.5 損失函數 50
3.6 強化學習優化與聯合訓練 51
3.6.1 理論-背景知識 52
3.6.2 Policy-Gradient 53
3.6.3 Monte-Carlo method 55
3.6.4 Self-critical Sequence Training 56
3.6.5 Joint Training 58
3.7 波束搜索法 58
第四章 實驗設計與結果 61
4.1 訓練及實驗環境 61
4.1.1 資料集 61
4.1.2 探索性資料分析 63
4.1.3 其他網路架構 66
4.1.4 參數設定和環境設置 67
4.2 評估工具 69
4.2.1 Rouge 70
4.2.2 BLEU 71
4.2.3 METEOR 72
4.3 實驗結果 75
4.3.1 關鍵詞萃取實驗 75
4.3.2 不同注意力機制下實驗 75
4.3.3 Extractive Summary實驗 85
4.3.4 Abstractive Summary實驗 86
4.3.5 聯合強化學習之Abstractive Summary實驗 91
4.3.6自行蒐集之評論摘要測試 93
第五章 結論與未來展望 97
5.1 結論 97
5.2 未來展望 97
參考文獻 98
附錄一、 傘繩/降落傘繩評論摘要結果 104
附錄二、 Prevue Hendryx 旅行鳥籠評論摘要結果 108
附錄三、 太陽能科學計算機評論摘要結果 111
附錄四、 單片式藥物測試組評論摘要結果 115
附錄五、 運動攝影機評論摘要結果 119
附錄六、 Note 8 Pro評論摘要結果 122
參考文獻 [1] Hu, M., & Liu, B. (2004, August). Mining and summarizing customer reviews. In Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 168-177).
[2] Khan, J., & Jeong, B. S. (2016, July). Summarizing customer review based on product feature and opinion. In 2016 international conference on machine learning and cybernetics (ICMLC) (Vol. 1, pp. 158-165). IEEE.
[3] Liu, Y., Li, X., & Wang, M. (2019, February). Quantifying Customer Review by Integrating Multiple Source of Knowledge. In Proceedings of the 2019 11th International Conference on Machine Learning and Computing (pp. 6-11).
[4] Somprasertsri, G., & Lalitrojwong, P. (2010). Mining Feature-Opinion in Online Customer Reviews for Opinion Summarization. J. UCS, 16(6), 938-955.
[5] Bengio, Y. (2008). Neural net language models. Scholarpedia, 3(1), 3881.
[6] Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S., & Dean, J. (2013). Distributed representations of words and phrases and their compositionality. In Advances in neural information processing systems (pp. 3111-3119).
[7] Pennington, J., Socher, R., & Manning, C. D. (2014, October). Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP) (pp. 1532-1543).
[8] Parkkinen, J., Selkäinaho, K., & Oja, E. (1990). Detecting texture periodicity from the cooccurrence matrix. Pattern Recognition Letters, 11(1), 43-50.
[9] Bojanowski, P., Grave, E., Joulin, A., & Mikolov, T. (2017). Enriching word vectors with subword information. Transactions of the Association for Computational Linguistics, 5, 135-146.
[10] Cavnar, W. B., & Trenkle, J. M. (1994, April). N-gram-based text categorization. In Proceedings of SDAIR-94, 3rd annual symposium on document analysis and information retrieval (Vol. 161175).
[11] Ng, K., & Zue, V. W. (2000). Subword-based approaches for spoken document retrieval. Speech Communication, 32(3), 157-186.
[12] Mani, I., & Maybury, M. T. (1999). Advances in Automatic Text Summarization. Cambridge.
[13] Jones, K. S. (2007). Automatic summarising: The state of the art. Information Processing & Management, 43(6), 1449-1481.
[14] Salton, G., & Buckley, C. (1988). Term-weighting approaches in automatic text retrieval. Information processing & management, 24(5), 513-523.
[15] Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent dirichlet allocation. Journal of machine Learning research, 3(Jan), 993-1022.
[16] Mihalcea, Rada, and Paul Tarau. "Textrank: Bringing order into text." Proceedings of the 2004 conference on empirical methods in natural language processing. 2004.
[17] Radev, D. R., Jing, H., Styś, M., & Tam, D. (2004). Centroid-based summarization of multiple documents. Information Processing & Management, 40(6), 919-938.
[18] Rossiello, G., Basile, P., & Semeraro, G. (2017, April). Centroid-based text summarization through compositionality of word embeddings. In Proceedings of the MultiLing 2017 Workshop on Summarization and Summary Evaluation Across Source Types and Genres (pp. 12-21).
[19] Erkan, Gunes, and Dragomir Radev. "Lexpagerank: Prestige in multi-document text summarization." Proceedings of the 2004 Conference on Empirical Methods in Natural Language Processing. 2004.
[20] Sutskever, I., Vinyals, O., & Le, Q. V. (2014). Sequence to sequence learning with neural networks. In Advances in neural information processing systems (pp. 3104-3112).
[21] Bahdanau, D., Cho, K., & Bengio, Y. (2014). Neural machine translation by jointly learning to align and translate. arXiv preprint arXiv:1409.0473.
[22] Hinton, G., Deng, L., Yu, D., Dahl, G. E., Mohamed, A. R., Jaitly, N., ... & Kingsbury, B. (2012). Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. IEEE Signal processing magazine, 29(6), 82-97.
[23] Rush, A. M., Chopra, S., & Weston, J. (2015). A neural attention model for abstractive sentence summarization. arXiv preprint arXiv:1509.00685.
[24] See, A., Liu, P. J., & Manning, C. D. (2017). Get to the point: Summarization with pointer-generator networks. arXiv preprint arXiv:1704.04368.
[25] Tu, Z., Lu, Z., Liu, Y., Liu, X., & Li, H. (2016). Modeling coverage for neural machine translation. arXiv preprint arXiv:1601.04811.
[26] Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., ... & Polosukhin, I. (2017). Attention is all you need. In Advances in neural information processing systems (pp. 5998-6008).
[27] Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.
[28] Kim, Y. H., Lewis, F. L., & Abdallah, C. T. (1997). A dynamic recurrent neural-network-based adaptive observer for a class of nonlinear systems. Automatica, 33(8), 1539-1543.
[29] Schuster, M., & Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE transactions on Signal Processing, 45(11), 2673-2681.
[30] Selvin, S., Vinayakumar, R., Gopalakrishnan, E. A., Menon, V. K., & Soman, K. P. (2017, September). Stock price prediction using LSTM, RNN and CNN-sliding window model. In 2017 international conference on advances in computing, communications and informatics (icacci) (pp. 1643-1647). IEEE.
[31] Graves, A., Mohamed, A. R., & Hinton, G. (2013, May). Speech recognition with deep recurrent neural networks. In 2013 IEEE international conference on acoustics, speech and signal processing (pp. 6645-6649). IEEE.
[32] Greff, K., Srivastava, R. K., Koutník, J., Steunebrink, B. R., & Schmidhuber, J. (2016). LSTM: A search space odyssey. IEEE transactions on neural networks and learning systems, 28(10), 2222-2232.
[33] Nallapati, R., Xiang, B., & Zhou, B. (2016). Sequence-to-sequence rnns for text summarization.
[34] Segaran, T., & Hammerbacher, J. (2009). Beautiful data: the stories behind elegant data solutions. " O'Reilly Media, Inc.".
[35] Ghahramani, Z. (2001). An introduction to hidden Markov models and Bayesian networks. In Hidden Markov models: applications in computer vision (pp. 9-41).
[36] Viterbi, A. (1967). Error bounds for convolutional codes and an asymptotically optimum decoding algorithm. IEEE transactions on Information Theory, 13(2), 260-269.
[37] Lew, R., & Mitton, R. (2013). Online English learners’ dictionaries and misspellings: One year on. International Journal of Lexicography, 26(2), 219-233.
[38] Peter Norvig. 2007. How to write a spelling corrector.
http://norvig.com/spell-correct.html. [Online; accessed 12-November-2016]
[39] Loria, S. (2018). TextBlob Documentation. Release 0.15.
[40] De Smedt, T., & Daelemans, W. (2012). Pattern for python. The Journal of Machine Learning Research, 13(1), 2063-2067.
[41] Min, W. N. S. W., & Zulkarnain, N. Z. (2020). Comparative Evaluation of Lexicons in Performing Sentiment Analysis. JOURNAL OF ADVANCED COMPUTING TECHNOLOGY AND APPLICATION (JACTA), 2(1), 14-20.
[42] Tesnière, L. (1959). Eléments de syntaxe structurale.
[43] 刘海涛. (2007). 泰尼埃的结构句法理论 (Doctoral dissertation).
[44] Abbasi Moghaddam, S. (2013). Aspect-based opinion mining in online reviews (Doctoral dissertation, Applied Sciences: School of Computing Science).
[45] Jiang, X., Hu, P., Hou, L., & Wang, X. (2018, August). Improving pointer-generator network with keywords information for chinese abstractive summarization. In CCF International Conference on Natural Language Processing and Chinese Computing (pp. 464-474). Springer, Cham.
[46] Paulus, R. (2019). U.S. Patent No. 10,474,709. Washington, DC: U.S. Patent and Trademark Office.
[47] A. Lamb, A. Goyal, Y. Zhang, S. Zhang, A. Courville, and Y. Bengio. Professor Forcing: A New Algorithm for Training Recurrent Networks (2016), NeurIPS 2016.
[48] S. Wiseman, and A. Rush. Sequence-to-Sequence Learning as Beam-Search Optimization (2016), EMNLP 2016.
[49] Rennie, S. J., Marcheret, E., Mroueh, Y., Ross, J., & Goel, V. (2017). Self-critical sequence training for image captioning. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 7008-7024).
[50] Yang, P., Ma, S., Zhang, Y., Lin, J., Su, Q., & Sun, X. (2018). A deep reinforced sequence-to-set model for multi-label text classification. arXiv preprint arXiv:1809.03118.
[51] Paulus, R., Xiong, C., & Socher, R. (2017). A deep reinforced model for abstractive summarization. arXiv preprint arXiv:1705.04304.
[52] Jianmo Ni. 2018. Amazon review data.
http://deepyeti.ucsd.edu/jianmo/amazon/index.html. [Online; accessed 12-November-2019]
[53] Boutkan, F., Ranzijn, J., Rau, D., & van der Wel, E. (2019). Point-less: More abstractive summarization with pointer-generator networks. arXiv preprint arXiv:1905.01975.
[54] Justifying recommendations using distantly-labeled reviews and fined-grained aspects, Jianmo Ni, Jiacheng Li, Julian McAuley
Empirical Methods in Natural Language Processing (EMNLP), 2019
[55] Ganapathibhotla, M., & Liu, B. (2008, August). Mining opinions in comparative sentences. In Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008) (pp. 241-248).
[56] Takase, S., & Okazaki, N. (2019). Positional encoding to control output sequence length. arXiv preprint arXiv:1904.07418.

[57] Shibata, Y., Kida, T., Fukamachi, S., Takeda, M., Shinohara, A., Shinohara, T., & Arikawa, S. (1999). Byte Pair encoding: A text compression scheme that accelerates pattern matching. Technical Report DOI-TR-161, Department of Informatics, Kyushu University.
[58] Liu, Y. (2019). Fine-tune BERT for extractive summarization. arXiv preprint arXiv:1903.10318.
[59] Brants, T. (2000, April). TnT: a statistical part-of-speech tagger. In Proceedings of the sixth conference on Applied natural language processing (pp. 224-231). Association for Computational Linguistics.
[60] Hunston, S. (2002). Pattern grammar, language teaching, and linguistic variation. Using corpora to explore linguistic variation, 167-183.
[61] Hoang, A., Bosselut, A., Celikyilmaz, A., & Choi, Y. (2019). Efficient adaptation of pretrained transformers for abstractive summarization. arXiv preprint arXiv:1906.00138.
[62] Lin, C. Y. (2004). Rouge: A package for automatic evaluation of summaries. In Text summarization branches out (pp. 74-81).
[63] Papineni, K., Roukos, S., Ward, T., & Zhu, W. J. (2002, July). BLEU: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting on association for computational linguistics (pp. 311-318). Association for Computational Linguistics.
[64] Miller, G. A. (1995). WordNet: A Lexical Database for English. Communications of the ACM, 38(11), 39-41. doi:10.1145/219717.219748
[65] Banerjee, S., & Lavie, A. (2005, June). METEOR: An automatic metric for MT evaluation with improved correlation with human judgments. In Proceedings of the acl workshop on intrinsic and extrinsic evaluation measures for machine translation and/or summarization (pp. 65-72).
[66] Maite Taboada, J. B. (2011, 06). Lexicon-Based Methods for Sentiment Analysis. Computational Linguistics, 37(2), 267-307. doi:10.1162/COLI_a_00049
[67] wikiwand.隱藏式馬可夫模型狀態變遷圖.(2020)
https://www.wikiwand.com/zh-tw/%E9%9A%90%E9%A9%AC%E5%B0%94%E5%8F%AF%E5%A4%AB%E6%A8%A1%E5%9E%8B
[68] He, R., & McAuley, J. (2016, April). Ups and downs: Modeling the visual evolution of fashion trends with one-class collaborative filtering. In proceedings of the 25th international conference on world wide web (pp. 507-517).
[69] McAuley, J., Targett, C., Shi, Q., & Van Den Hengel, A. (2015, August). Image-based recommendations on styles and substitutes. In Proceedings of the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval (pp. 43-52).
[70] 小風英文教室- 連讀與縮寫表(2005)
https://iter01.com/372682.html
[71] 英文縮寫點的用法 & 常見的英文縮寫(2014) https://aixo.pixnet.net/blog/post/43232002
[72] wikiwand.維特比演算法. (2020)
https://www.wikiwand.com/zh-tw/%E7%BB%B4%E7%89%B9%E6%AF%94%E7%AE%97%E6%B3%95#/%E6%B3%A8%E9%87%8A.
[73] Forney Jr, G. D. (2005). The viterbi algorithm: A personal history. arXiv preprint cs/0504020.
[74] Fan and Fuel研究調查(2017)
https://fanandfuel.com/no-online-customer-reviews-means-big-problems-2017/
[75] MarketPlace研究調查(2016)
https://www.marketplacepulse.com/articles/how-many-retailers-are-selling-on-amazon-marketplaces
[76] Episerver Ascend研究調查(2019)
https://retailwire.com/discussion/who-is-winning-the-shopping-search-race-amazon-or-google/
[77] Raymond James研究調查(2014)
https://www.businessinsider.com/amazon-gains-share-in-ecommerce-searches-2017-1
[78] Julian McAuley研究調查(2014)
https://minimaxir.com/2014/06/reviewing-reviews/
論文全文使用權限
  • 同意授權校內瀏覽/列印電子全文服務,於2020-09-01起公開。
  • 同意授權校外瀏覽/列印電子全文服務,於2020-09-01起公開。


  • 如您有疑問,請聯絡圖書館
    聯絡電話:(06)2757575#65773
    聯絡E-mail:etds@email.ncku.edu.tw