引導式測驗 · 中等 等級

人工智慧讀者諮詢測驗

探索人工智慧如何推薦書籍,為什麼產生的標題和描述需要目錄驗證,以及圖書館如何保護隱私和多樣化的發現。

相關引導路徑AI 讀者諮詢書籍推薦
問題 1 的 8

What can an AI readers’ advisory tool usefully do?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. What can an AI readers’ advisory tool usefully do?
    • Suggest titles based on stated preferences for readers to explore
    • Guarantee every suggested book exists
    • Decide what a reader should enjoy
    • Replace library catalog verification
  2. How should readers verify a generated title?
    • Assume the author name is accurate
    • Ask the same model to confirm its list
    • Check a trusted library catalog or publisher source
    • Search only for similar-sounding titles
  3. Why preserve non-personalized browsing paths?
    • They guarantee every book appears equally
    • They let readers explore without depending on a personal profile
    • They eliminate the need for metadata
    • They prevent librarians from making suggestions
  4. What privacy risk can personalization create?
    • Book titles cannot be linked to people
    • Personalized ranking requires no data
    • Reading histories are always public
    • Reading interests and histories may be sensitive
  5. Which inference should a system avoid treating as certain based on a reader profile alone?
    • The system knows the reader’s motivation or taste with certainty
    • The catalog lists the title under a recorded genre
    • The reader explicitly selected a preference in this session
    • The library catalog records that the title is in its collection
  6. Which quality measure extends beyond click counts?
    • Only the number of recommendations returned
    • Only system response speed
    • Relevance, novelty, diversity, and catalog coverage
    • The number of profile fields collected
  7. What can cause a recommender to narrow discovery?
    • Offering filters and varied browsing paths
    • Repeatedly ranking popular titles and familiar genres
    • Checking title records in the catalog
    • Inviting corrections from readers
  8. How can librarians support a reader beyond an automated list?
    • Automatic confirmation of every model output
    • A guarantee that every book will be available
    • An obligation to follow algorithmic rankings
    • Contextual help interpreting preferences and collection knowledge