[AI‑Powered Semantic Search — public beta]
CBETA Online「語意搜尋」功能開放測試!
法鼓文理學院結合 OpenAI API 與 RAG(Retrieval‑Augmented Generation)技術,首次將生成式 AI 應用於 CBETA 資料庫,為線上版 CBETA Online 帶來不同的經文檢索體驗。
與傳統關鍵詞搜尋不同,您只需輸入一行問題或主題,系統即以語意向量比對,找出最相關段落,整合重點、引經據典,並提出延伸議題。所對比的經文依相關度排序,連接 CBETA Online 原文。
搜尋範例:
目前可鎖定「阿含部」、「律部」等部類;未來或可另指定藏經、叢書等。詳見下方〈操作引導〉。
歡迎試用並回饋寶貴意見!
P.S.: 更多詳情,請線上參與 5/22 (四) 2025 法鼓數典專案春季成果發表會,CBETA 執行長洪振洲教授將親自分享本實驗與後續規劃。
CBETA Online now offers a public beta of its experimental feature Semantic Search!
Developed by the Dharma Drum Institute of Liberal Arts using OpenAI API and Retrieval Augmented Generation (RAG) technology, this marks the team’s first attempt to integrate generative AI into the CBETA database.
Unlike traditional keyword search, one simply enters a question or topic. The engine compares the topic to vectorised CBETA texts, retrieves the most relevant passages, and generates an overview with references and follow-up questions. Text references are also ranked by relevance and link directly to the texts in CBETA Online.
At the moment one may filter by thematic groups (e.g., Āgama, Vinaya), with more options—canons, series, and beyond—planned soon. See Quick Guide below for details.
We invite you to explore this feature and would love to hear from you!
P.S.: For a deeper dive, join us online on Thursday, May 22 for the 2025 Spring Showcase of Dharma Drum Digital Archive Projects (法鼓數典專案春季成果發表會). Prof. Jenjou Hung, CBETA CEO, will share firsthand updates on this experiment and its future roadmap.
操作引導 Quick Guide
1. 開啟「語意搜尋」Launch Semantic Search
2. 輸入問題 Enter a query
3. 搜尋結果 Result
綜合說明:AI 整合前 20 筆相關經文的重點。
Summary: AI synthesizes key points from the 20 most relevant passages.
4. 意見回饋 Feedback
點一下右下角的意見回饋泡泡,讓我們知道您的使用心得!
Click the feedback bubble in the bottom-right corner and let us know how it worked for you!