Technical Name The Dialogue Semantic Analysis of Complex Tasks and Deep Consumption Needs
Project Operator National Cheng Kung University
Project Host
Summary
For the shortcomings of the traditional question answering systems can only handle single question intent analysis, this “Dialogue Semantic Analysis of Complex Tasks and Deep Consumption Needs” has developed a practical model for dialogue semantic analysis of complex task and deep consumption need as the basic analysis framework, which can solve ellipsis problem of key semantic structure of contextual questions. When the latter question lacks a main target focus, this model can automatically generate the key semantic structure based on the previous questions, and then obtain the correct semantic structure of the complete intent of the contextual questions.
Scientific Breakthrough
針對傳統問答系統僅能處理單一問句意圖分析的缺點,發展複雜任務及深層消費需求之對話語意分析模型,解決關聯問句關鍵語意結構省略問題。當問句缺乏關鍵詞彙時,自動增補關鍵語意結構,進行意圖的正確語意結構分析。
Industrial Applicability
 我們所開發出的任務型生活服務聊天機器人,主要是以『人、生活事件、複雜任務、消費需求、商品服務』為導向,開發出更人性化的自然語言互動介面,提供食衣住行育樂醫療各個層面生活服務,扼要簡述幾個實用服務亮點如下:
1. 醫療:避免老人亂服成藥,當使用者來告訴聊天機器人,目前身體有哪些症狀,例如:咳嗽、頭痛、頻尿等狀況,機器人會協助使用者判斷可能罹患什麼疾病,進而建議該前往哪一個科別就診。
2. 旅宿:旅行對於現代人來說,已經成為一個放鬆心靈的解藥,我們提供了幾個相關服務,例如:飯店預訂、景點餐廳查詢、詢問天氣、高鐵班次查詢。
3. 休閒育樂:其他基本生活服務,包括:購買書籍、購買商品、預購演唱會及球賽門票、網路音樂及影片欣賞。
Keyword
Notes
  • Contact
other people also saw