Skip navigation

DSpace JSPUI

DSpace preserves and enables easy and open access to all types of digital content including text, images, moving images, mpegs and data sets

Learn More
DSpace logo
English
中文
  • Browse
    • Communities
      & Collections
    • Publication Year
    • Author
    • Title
    • Subject
    • Advisor
  • Search TDR
  • Rights Q&A
    • My Page
    • Receive email
      updates
    • Edit Profile
  1. NTU Theses and Dissertations Repository
  2. 電機資訊學院
  3. 資訊網路與多媒體研究所
Please use this identifier to cite or link to this item: http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/87595
Title: 基於記憶庫問答的記憶輔助機器人
Memory Assistance Robot based on Memory Bank Question Answering
Authors: 宋豐裕
Feng-Yu Sung
Advisor: 傅立成
Li-Chen Fu
Keyword: 記憶輔助機器人,記憶輔助,開放領域問答,問答系統,
Memory Assistance Robot,Memory Assistance,Open-Domain Question Answering,Question Answering,
Publication Year : 2023
Degree: 碩士
Abstract: 在日常生活,人們需要記憶各式各樣的事情,也時常需要作筆記,避免日後遺忘。因此,本論文提出一個基於記憶庫問答的記憶輔助機器人,主要以語音作為輸入,幫助使用者儲存記憶和回憶。
我們所提出系統有多個模組,包含記憶儲存模組、記憶庫問答模組和記憶日曆模組。記憶儲存模組能讓使用者透過語音來儲存記憶,我們的系統會通過語音辨識來得到逐字稿,接著使用多種方法抽取特徵,建立記憶並存儲在記憶庫中。而記憶庫問答模組,目的是讓使用者透過問答的方式回憶。根據問題,我們基於開放領域問答的架構來進行檢索與預測,接著使用一個答案排序的方法來得到最終的答案。除此之外,為了幫助使用者瀏覽記憶,我們提出一個記憶日曆模組,從逐字稿中提取日常事件,並顯示在日曆上。實驗部分,我們使用兩個實驗進行評估,第一個實驗為開放領域問答的實驗,而第二個實驗則是人機互動實驗。
In daily life, people need to memorize various things and often need to write notes to avoid forgetting in the future. Therefore, this thesis proposes a memory assistance robot based on memory bank question answering, which mainly uses speech as input to help the user store and recall memories.
Our proposed system contains multiple modules, including memory storage module, memory bank question answering module, and memory calendar module. The memory storage module allows users to store memories through speech. Our system will obtain transcript through speech recognition and then use various methods to extract features, construct the memory, and store it in the memory bank. For the memory bank question answering module, the purpose is to allow the user to recall through question answering. According to the query, we perform retrieval and prediction based on the open-domain question answering architecture. Then, we use an answer ranking method to get the final answer. In addition, to help the user view memories, we propose a memory calendar module, which extracts daily events from the transcript and shows them on the calendar. In the experiment part, we use two experiments for evaluation. The first experiment is the open-domain question answering experiment, and the second experiment is the human-robot interaction experiment.
URI: http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/87595
DOI: 10.6342/NTU202300222
Fulltext Rights: 未授權
Appears in Collections:資訊網路與多媒體研究所

Files in This Item:
File SizeFormat 
ntu-111-1.pdf
  Restricted Access
5.79 MBAdobe PDF
Show full item record


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

社群連結
聯絡資訊
10617臺北市大安區羅斯福路四段1號
No.1 Sec.4, Roosevelt Rd., Taipei, Taiwan, R.O.C. 106
Tel: (02)33662353
Email: ntuetds@ntu.edu.tw
意見箱
相關連結
館藏目錄
國內圖書館整合查詢 MetaCat
臺大學術典藏 NTU Scholars
臺大圖書館數位典藏館
本站聲明
© NTU Library All Rights Reserved