中大機構典藏-NCU Institutional Repository-提供博碩士論文、考古題、期刊論文、研究計畫等下載:Item 987654321/93356
English  |  正體中文  |  简体中文  |  Items with full text/Total items : 80990/80990 (100%)
Visitors : 41692061      Online Users : 1470
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version


    Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/93356


    Title: 基於大型語言模型應用指示詞打造無程式碼對話系統平台 - 以聊故事機器人為例;Prompts are All your Need to Construct a Dialogue System - A Case Study on Story Chatbots
    Authors: 林子平;Lin, Tzu-Ping
    Contributors: 資訊工程學系
    Keywords: 對話系統;對話代理人;指示學習;無程式碼人工智慧;Dialogue System;Conversational Agent;Instruction Learning;No-Code Artificial Intelligence
    Date: 2023-07-28
    Issue Date: 2024-09-19 16:55:38 (UTC+8)
    Publisher: 國立中央大學
    Abstract: 過去的教育類型對話代理人存在兩個主要難點。首先,在語言理解能力有限的情況下,往往只能採取主導式的問答模式。這種模式下,對話代理人通常只能提供選項供學生選擇回答,而即使允許學生自由回答,回應通常也僅限於樣板式的回覆。這嚴重限制了對話系統的互動性和個性化,對教學成效造成了影響。其次,針對不同程度、背景的學生所採取的教學策略和方法通常需要有相對應的教學理論支持。然而,開發這樣的系統非常耗時且費工,往往僅能實作少數教學策略。
    為了解決這個問題,我們建立了一個名為Educational Agent Crafting Tool(簡稱EduACT)的教育對話代理人創造平台。這個平台利用強大的語言模型ChatGPT,幫助教育工作者創造教學助教。與被動等待學生發問不同,我們採用了自然語言指示(Prompt)引導對話代理人,提供主動對話代理人類型設定。這使得教育工作者可以根據不同的教學內容和對象快速建立多個對話代理人,從而提供學生的一對一教學的可能。EduACT 平台的核心在於解構教學者與學生之間的對話方式,使其能夠以自然的方式與學生互動,回答問題、提供資訊並執行特定任務,從而增加教學的互動性。為了讓沒有程式背景的教育工作者容易使用,我們提出了一個模組化架構方法。創作者可以為對話代理人設計多個模組,每個模組負責不同的動作任務。這樣的設計使得對話過程比直接使用ChatGPT 更加主動和可控,處理多輪對話時回應更加一致合理,避免對話代理人提供不適當的回答,或是被學生引導至與課程無關的話題。
    我們使用EduACT 平台設計了三個教師類型的對話代理人「魚姊姊」家族,以及三個學生類型的「熊小弟」們。我們讓兩者進行對話,並且顯示了它們能夠成功扮演各自角色,並且使用設定的任務模組進行完整的對話。總結來說,EduACT 平台提供了有教育專業背景但是沒有程式基礎的教育工作者直觀且容易上手的界面,讓他們能夠建立多元且高互動性的對話代理人,以確保對話代理人能夠更好地融入教學場景,提供更有效的教學輔助和學習體驗。;In the past, educational conversational agents faced two main challenges. Firstly, their limited language understanding resulted in a predominantly directive question-and-answer format. This constrained interactivity and personalization, affecting teaching effectiveness. Secondly, creating tailored teaching strategies for students with different levels and backgrounds was time-consuming and challenging.
    To tackle these issues, we developed Educational Agent Crafting Tool (EduACT), a platform that utilizes ChatGPT to help educators create teaching assistants. Instead of passively waiting for questions, EduACT uses natural language prompts to guide active conversations, enabling educators to quickly create multiple agents for personalized one-on-one instruction.EduACT deconstructs dialogue between educators and students, allowing natural interactions for answering questions, providing information, and executing tasks, thus increasing teaching interactivity. To simplify usage for educators without programming knowledge, we introduced a modular approach. This ensures more controlled and consistent responses during multi-turn conversations, avoiding irrelevant topics.
    Using EduACT, we designed three teacher types “Fish Sisters” and three student types “Bear Brothers” for successful role portrayal and comprehensive conversations. Overall, EduACT empowers educators to create diverse, interactive conversational agents, enhancing teaching assistance and learning experiences.
    Appears in Collections:[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

    Files in This Item:

    File Description SizeFormat
    index.html0KbHTML15View/Open


    All items in NCUIR are protected by copyright, with all rights reserved.

    社群 sharing

    ::: Copyright National Central University. | 國立中央大學圖書館版權所有 | 收藏本站 | 設為首頁 | 最佳瀏覽畫面: 1024*768 | 建站日期:8-24-2009 :::
    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - 隱私權政策聲明