隨著深度學習和自然語言處理技術的進步，任務導向聊天機器人廣泛地被應用到各種領域，協助使用者解決問題，並提供了更便利的方式與電腦進行互動。Rasa是一個用來建置任務導向聊天機器人之開源的對話AI框架。本論文提出一個協助性框架，目的是為了來減輕使用Rasa建置任務導向聊天機器人之開發者的負擔。我們從任務導向聊天機器人的功能面和風格面，在功能方面，我們設計了一套程序來指定句子的意圖並識別關鍵實體，然後使用滿足需求的原始對話，來生成在Rasa中的對話管理模型所需的訓練數據(故事)；在風格面，我們提出了三種不同的風格特徵(個性、易讀性、情感)，以幫助聊天機器人開發者去驗證大量對話數據。此步驟保留了滿足開發者需求之風格的原始對話來構建聊天機器人，讓聊天機器人可以以理想的風格進行回覆，並塑造聊天機器人的形象。
最後，我們根據我們提出的流程，建立了一個旅遊領域的聊天機器人，可以用來預訂班機和找尋飯店。

With the advance of deep learning and natural language processing technology, task-oriented chatbots have been widely applied in many fields to help users solve problems and provide a more convenient way to interact with computers. Rasa is an open-source conversational AI framework to build task-oriented chatbots. This research proposed a supportive framework to reduce the effort of developers who use Rasa to build task-oriented chatbots. This framework help users more efficiently address issues in both the functional aspect and the style aspect when designing task-oriented chatbots. In the functional aspect, we design a set of methods to identify intent and identify key entities from a sentence, which are needed to build the dialogue management model in Rasa. These methods are learned from the original dialogues which serve as the training data (stories). In the style aspect, we propose three different style features (personality, readability, emotion) to help chatbot developers verify if the dialogue data which will be used as training data are indeed exhibit the desired styles. The purpose is to construct a chatbot that can respond in the ideal style and shape the image of chatbot. Finally, we evaluate the various components in our framework and develop a travel chatbot on the tourism field by applying our proposed framework on real flight/hotel booking data.

論文審定書 i
誌謝 ii
摘要 iii
Abstract iv
Table of Contents vi
List of Figures vii
List of Tables viii
Chapter 1－ Introduction 1
1.1 Background and Motivation 1
1.2 Thesis Organization 4
Chapter 2－ Related Work 5
2.1 Construction of task-oriented chatbots 5
2.2 Rasa 9
2.3 Style of chatbot 12
Chapter 3－ Research Method 15
3.1 Architecture of Our Approach 15
3.2 Style Feature Derivation 16
3.2.1 Personality 16
3.2.2 Readability 18
3.2.3 Emotion 21
3.3 Functional suggestion 22
3.3.1 Sentence Labeling 24
3.3.2 Entity Annotation and Word Extraction 25
3.3.3 Intent and Action Classification 26
3.3.4 Story Formatting 27
Chapter 4－ Prototyping and Evaluation 30
4.1 Prototyping 30
4.1.1 Data Description 30
4.1.2 Style Feature Derivation 31
4.1.3 Functional Suggestion 35
4.1.4 Prototype of Chatbot 39
4.2 Evaluation 41
4.2.1 Intent and action classification 41
Chapter 5－ Conclusion 46
References 47

Su, P.-H., Gašić, M., Mrkšić, N., Rojas-Barahona, L. M., Ultes, S., Vandyke, D., . . . Young, S. (2016). On-line active reward learning for policy optimisation in spoken dialogue systems.
Young, S., Gašić, M., Thomson, B., & Williams, J. D. (2013). Pomdp-based statistical spoken dialog systems: A review. Proceedings of the IEEE, 101(5), 1160-1179.
Oriol Vinyals, & Le, Q. (2015). A neural conversational model.
Chen, H., Liu, X., Yin, D., & Tang, J. (2017). A survey on dialogue systems: Recent advances and new frontiers. Paper presented at the ACM SIGKDD Explorations Newsletter.
Peng, Z. H., & Ma, X. J. (2019). A survey on construction and enhancement methods in service chatbots design. Transactions on Pervasive Computing Interaction, 1-20.
Weston, J., Chopra, S., & Bordes, A. (2014). Memory networks. Paper presented at the ICLR 2015.
Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780.
Schuster, M., & Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE transactions on Signal Processing, 45(11), 2673-2681.
Bocklisch, T., Faulkner, J., Pawlowski, N., & Nichol, A. (2017). Rasa: Open source language understanding and dialogue management.
Williams, J. D., Asadi, K., & Zweig, G. (2017). Hybrid code networks: practical and efficient end-to-end dialog control with supervised and reinforcement learning.
Nikitina, S., Callaioli, S., & Baez, M. (2018). Smart conversational agents for reminiscence. Paper presented at the Proceedings of the 1st International Workshop on Software Engineering for Cognitive Services.
Hung, V., Elvir, M., Gonzalez, A., & DeMara, R. (2009). Towards a method for evaluating naturalness in conversational dialog systems. Paper presented at the 2009 IEEE International Conference on Systems, Man and Cybernetics.
Rebecca J. Passonneau, Alvarado, I., Crone, P., & Jerome, S. (2011). PARADISE-style evaluation of a human-human library corpus. Paper presented at the Proceedings of the SIGDIAL 2011 Conference.
Kuligowska, K. (2015). Commercial chatbot: Performance evaluation, usability metrics and quality standards of embodied conversational agents. Paper presented at the Professionals Center for Business Research.
Radziwill, N. M., & Benton, M. C. (2017). Evaluating quality of chatbots and intelligent conversational agents.
Tausczik, Y. R., & Pennebaker, J. W. (2010). The psychological meaning of words: LIWC and computerized text analysis methods. Journal of language social psychology, 29(1), 24-54.
Jongerius, C. (2018). Quantifying Chatbot Performance by using Data Analytics. (Master Thesis). Utrecht University,
Angeli, A. D., Johnson, G. I., & Coventry, L. (2001). The unfriendly user: exploring social reactions to chatterbots. Paper presented at the Proceedings of The International Conference on Affective Human Factors Design, London.
Xiao, H., Reid, D., Marriott, A., & Gulland, E. K. (2005). An adaptive personality model for ECAs. Paper presented at the International Conference on Affective Computing and Intelligent Interaction.
Callejas, Z., López-Cózar, R., Ábalos, N., & Griol, D. (2011). Affective conversational agents: the role of personality and emotion in spoken interactions: IGI Global.
Norman, W. T. (1963). Toward an adequate taxonomy of personality attributes: Replicated factor structure in peer nomination personality ratings. The Journal of Abnormal and Social Psychology, 66(6), 574.
McCrae, R. R., & Jr., P. T. C. (1985). The NEO personality inventory.
Haan, H. d. (2018). Chatbot Personality and Customer Satisfaction. (Master Thesis). Utrecht University,
Smestad, T. L. (2018). Personality Matters! Improving The User Experience of Chatbot Interfaces-Personality provides a stable pattern to guide the design and behaviour of conversational agents. (Master Thesis). NTNU,
Flesh, R. (1948). A new readability yardstick. Journal of Applied Psychology, 32(3), 221.
Gunning, R. (1952). The Technique of Clear Writing. New York: McGraw-Hill.
McLaughlin, G. H. (1969). SMOG grading-a new readability formula. Journal of Reading, 12(8), 639-646.
Kincaid, J. P., Jr., R. P. F., Rogers, R. L., & Chissom, B. S. (1975). Derivation of new readability formulas (automated readability index, fog count and flesch reading ease formula) for navy enlisted personnel.
Dale, E., & Chall, J. S. (1995). Readability Revisted: The New Dale-Chall Readability Formula. Cambridge: Brookline Books.
Senter, R. J., & Smith, E. A. (1967). Automated readability index. Retrieved from
荊溪昱. (1995). 中學國文教材的適讀性研究: 適讀年齡值的推估.
宋曜廷, 陳茹玲, 李宜憲, 查日龢, 曾厚強, 林維駿, . . . 張國恩. (2013). 中文文本可讀性探討: 指標選取, 模型建立與效度驗證. Paper presented at the 中華心理學刊.
Sung, Y. T., Chang, T. H., Lin, W. C., Hsieh, K. S., & Chang, K. E. (2016). CRIE: An automated analyzer for Chinese texts. Behavior research methods, 48(4), 1238-1251.
Griol, D., Molina, J. M., & Callejas, Z. (2015). Towards emotionally sensitive conversational interfaces for E-therapy. Paper presented at the International Work-Conference on the Interplay Between Natural and Artificial Computation.
Lemon, O., & Pietquin, O. (2012). Data-driven methods for adaptive spoken dialogue systems: Computational learning for conversational interfaces: Springer Science & Business Media.
Mohammad, S. M., & Turney, P. D. (2013). Crowdsourcing a word–emotion association lexicon. Computational Intelligence, 29(3), 436-465.
Manning, C., Surdeanu, M., Bauer, J., Finkel, J., Bethard, S., & McClosky, D. (2014). The Stanford CoreNLP natural language processing toolkit.
Honnibal, M., & Montani, I. (2017). spacy 2: Natural language understanding with bloom embeddings, convolutional neural networks and incremental parsing. To appear, 7(1).
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. Paper presented at the arXiv preprint arXiv:1810.04805.
Asri, L. E., Schulz, H., Sharma, S., Zumer, J., Harris, J., Fine, E., . . . Suleman, K. (2017). Frames: A corpus for adding memory to goal-oriented dialogue systems. Paper presented at the Proceedings of the 18th Annual SIGdial Meeting on Discourse and Dialogue
Bunk, T., Varshneya, D., Vlasov, V., & Nichol, A. (2020). DIET: Lightweight Language Understanding for Dialogue Systems. Paper presented at the arXiv preprint arXiv:2004.09936.