近年來，隨著人工智慧的普及，聊天機器人應用程式的發展也越來越快。如今，許多業者在社群平台上都擁有自己的聊天機器人，但是這些聊天機器人中的大多數只能處理一些常規任務，以餐飲業為例，他們的聊天機器人只能詢問用餐內容，餐廳訊息或進行預訂。但是，借助自然語言處理技術，能夠真正理解人類語言和需求的聊天機器人是我們所期待的。

在這篇論文中，我們希望可以建立一個飲食推薦系統，該系統可以在決定吃什麼或點菜時了解人類的需求，然後通過該系統建立一個多輪對話式聊天機器人。我們定義三種類型的使用者需求：疾病和症狀，心情和特殊需求。我們的系統將建立多個模型並收集多種資源，以在與使用者對話期間抽取這些使用者需求。最後，我們將根據所抽取出的身心和飲食需求向使用者推薦合適的食物清單。

In recent years, as the artificial intelligent become very popular, the development of chatbot application has also become faster. Nowadays, many stores have there own chatbot on the social community, but most of those chatbot can only handle some regular task, taking the catering as example, their chatbots can only asking the content of meals, information of the restaurant or making reservation. But with the Natural language processing skills, chatbots that can truly understand human’s language and needs are what we look forward to.

In our work, we hope we can build a diet recommendation system that can understand human’s needs when deciding what to eat or ordering food and then build a multi-round conversational chatbot by the system. We define three type of user needs: Disease & Symptom, emotion, and Special need. Our system will build multiple model and collect multiple resource to extract these user needs during the conversation with user. In the end, we will recommend a suitable food list to users according to the extracted physical, mental, and dietary needs.

[1] Jhih-Sheng Fan, Chatbot Application: Event driven Task-Oriented Store Recommendation.
[2] Wen-Chun Wu, A Health-preserving Suggestions Chatbot based on Disease Knowledge-graph.
[3] Jia-Chun Cai, Summary Generation for Chinese Patient Complaint based on Medical Entity Recognition and Medical Terminology Mapping.
[4] Vaswani et al., Attention Is All You Need.
[5] Gaël Letarte, et al., Importance of Self-Attention for Sentiment Analysis.
[6] Sheng-Tang Chuang, Product Recommendation Bot Based On Related Task Structure Using News Event and Yahoo Answer.
[7] Chen-An Wang, Shopping Chatbot based on Complex Task Structure and Consumption Need.
[8] Ting-Xuan Wang, Identifying Consumption Needs for Complex Search Tasks to Recommend Proper Advertisements.
[9] Zhiheng Huang, Wei Xu, Kai Yu. Bidirectional LSTM-CRF Models for Sequence Tagging
[10] Laura Canetti, Eytan Bachar. Elliot M. Berry, Food and emotion.
[11] Jorge Almeida. Personalized Food Recommendations Exploring Content-Based Methods.
[12] Burusothman Ahiladas, Paraneetharan Saravanaperumal, Sanjith Balachandran, Thamayanthy Sripalan and Surangika Ranathunga. Ruchi: Rating Individual Food Items in Restaurant Reviews.
[13] Tomas Mikolov, et al.. Distributed Representations of Words and Phrases and their Compositionality.
[14] John Lafferty, Andrew McCallum, Fernando C.N. Pereira. Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data.
[15] Sepp Hochreiter, Jürgen Schmidhuber. Long Short-Term Memory.