日常生活中經常碰見聊天時需要對個人過去生活經歷進行回憶的情況,且不是每次都能正確快速的回憶起過去經歷的細節,常常只能回憶起部分的內容,只差一點就能完整想起。基於此,本論文定義出一主動式的系統,根據使用者在言談中透露的資訊,在使用者進入回想的狀況時,主動搜尋出對應的過去生活經歷。本論文定義此使用情境,將其轉化成對於生活情境的對話中,對於生活紀錄檢索的問題,再依照定義去搜集資料,並進行實驗。因為過去並無類似的資料集,故同時分析了生活情境對話的要素,並設計一套流程化方法以搜集包含個人生活紀錄與隱含生活紀錄之言論,再依此流程搜集資料進行實驗。本論文以資料檢索之方法,根據生活情境對話找出對應之生活紀錄,並提供所缺失的回憶提供給使用者。
In daily life, we often recall things in the past from memory in casual conversation, but not always rapidly recall it correctly. Sometimes we even barely recall it but fail. Base on this scenario, this thesis defines a proactive system, which according to information in utterance of user, retrieves and provides the information the user needs when he/she is recalling from memory. This thesis defines the scenario, converts into a problem of life event retrieval in casual conversation. And follow the definition to collect data and perform experiment. Since this kind of dataset not exists, we analysis the key elements in the scenario and design a workflow to collect life events and also the utterance related to the events. We retrieve life events according to casual conversation, and provide the missing memory to the user.