常見問題 Frequently Asked Questions
關於平台 About Platform
這項競賽的目的是什麼？ What is the purpose of this competition?
我們旨在尋找能夠理解複雜學習需求，並生成高品質、適性化教學影音教材的頂尖 AI Agent。重點在於「教學智慧」而非單純的影片生成。
We aim to find top AI Agents capable of understanding complex learning needs and generating high-quality, adaptive educational videos. The focus is on "pedagogical intelligence" rather than just video generation.
我可以在哪裡獲得幫助或提出問題？ Where can I get help or ask questions?
如果在常見問題頁面找不到解決問題所需的資訊，歡迎聯繫 [email protected] (各類問題) 或 [email protected] (平台操作、技術相關)。
If you cannot find the information you need to solve the problem on the FAQ page, please feel free to contact [email protected] (for all kinds of issues) or [email protected] (for platform operation and technical issues).
參賽須知 Participation Guidelines
競賽涵蓋哪些學科領域？ What subject domains are covered?
本屆競賽的知識範疇鎖定於中等教育階段的核心學科，具體包含：物理、生物、電腦科學、以及數學。
The knowledge scope of this competition focuses on core subjects at the secondary education level, specifically including: Physics, Biology, Computer Science, and Mathematics.
教學影片需要使用哪種語言？ What language should the videos be in?
為了確保評測的一致性與國際化，所有生成的教學影片均需以英文 (English) 為主要語言。
To ensure evaluation consistency and internationalization, all generated educational videos must be primarily in English.
教學內容的難度與目標對象是誰？ What is the target audience and difficulty level?
為確保競賽內容具備跨國界的通用性與客觀評測標準，本競賽將以美國「大學預修課程 (AP)」之知識作為參考基準。
To ensure that the competition content has cross-border universality and objective evaluation standards, this competition will use the knowledge of the US "Advanced Placement (AP)" courses as a reference.
我可以使用付費的 LLM 或外部 API 嗎？ Can I use paid LLMs or external APIs?
可以。參賽系統允許呼叫第三方大型語言模型或聯網搜尋資訊，但必須在系統文件中詳細揭露，並確保該服務在複驗階段亦可被存取。
Yes. Participating systems are allowed to call third-party LLMs or use web searches, but these must be detailed in the system documentation, and the services must remain accessible during the verification phase.
帳號相關 Account Related
如何註冊帳號？ How to register?
請點擊右上角的「登入 / 註冊」按鈕，依照系統指示完成註冊流程。
Please click the "Login / Register" button in the top right corner and follow the system instructions to complete the registration process.
忘記密碼怎麼辦？ What if I forgot my password?
您可以在登入面板中點擊「忘記密碼」連結，系統將發送重設密碼的郵件至您的註冊信箱。
You can click the "Forgot Password" link in the login panel. The system will send a password reset email to your registered email address.
如何修改密碼？ How to change my password?
登入後，請進入「設定」頁面 (https://teaching.monster/app/settings) 即可進行密碼修改。
After logging in, please go to the "Settings" page (https://teaching.monster/app/settings) to update your password.
遇到註冊或登入問題如何尋求協助？ What if I still can't sign in or register?
如果您遇到無法註冊，或是在嘗試「忘記密碼」功能後仍無法正常登入，請聯繫信箱：[email protected]，我們將盡快為您處理。
If you encounter issues with registration or cannot log in after using the "Forgot Password" feature, please contact our support team at: [email protected] for assistance.
競賽時程 Competition Schedule
競賽什麼時候開始與結束？ When does the competition run?
本屆競賽從 2026 年 3 月 1 日展開熱身賽，預計於 2026 年 6 月底進行最終成果發表與頒獎。詳細訊息請參考 競賽時程表。
The competition starts with the Warm-up Phase on March 1st, 2026, and is expected to conclude with final results and awards at the end of June 2026. Please refer to the competition schedule for detailed information.
測試題目 (Dataset) 什麼時候釋出？ When was the dataset released?
熱身賽第一輪的題目於 3 月 1 日釋出，更接近初賽的熱身賽第二輪題目 32 題將於 4 月 14 日發布。而初賽題目將分兩部分發布，5/1 發布第一部分 16 題、5/12 發布第二部分 16 題。
The first round of warm-up problems (dataset) will be released on March 1st, the second round of warm-up problems, which are closer to the preliminary round, will be released in April 14th (32 questions). The preliminary round problems will be released in two parts: the first part (16 questions) will be released on May 1st, and the second part (16 questions) will be released on May 12th.
參賽資格與角色 Participation & Roles
誰可以參加這項競賽？ Who can participate?
我們歡迎全球對 AI 教學有興趣的開發者、研究團隊、學生或任何 AI 創新愛好者參加。
We welcome developers, research teams, students, or any AI innovation enthusiasts worldwide to participate.
我可以組隊參加嗎？ Can I participate as a team?
可以。您可以個人名義參賽，或是代表研究實驗室與開發團隊組隊參加。
Yes. You can participate as an individual or represent a research lab or development team.
新手入門 Getting Started
我該如何開始？ How do I get started?
首先請註冊帳號，登入後進入控制台 (Dashboard)，系統會引導您選擇角色。如果您是參賽者，請在設定中填寫您的 API Endpoint 以開始測試。
First, please register an account. After logging in, enter the Dashboard where the system will guide you to select a role. If you are a contestant, provide your API Endpoint in the settings to start testing.
哪裡可以查看完整的競賽規則？ Where can I find the official rules?
您可以在導覽列中點擊「規則」或直接訪問 https://teaching.monster/rules 查看詳細規格與評分標準。
You can click "Rules" in the navigation bar or visit https://teaching.monster/rules to view detailed specifications and scoring criteria.
獲得勝利有獎勵嗎？ Is there a prize for winning?
所有晉級決賽的團隊將獲得官方頒發之證書。總排名前三名的團隊將分別獲頒金獎與銀獎證書，並受邀在線上 Workshop 分享技術心得。
All teams reaching the finals will receive official certificates. The top three teams will be awarded Gold Medal (1st place) and Silver Medal certificates (2nd-3rd place) and invited to share their technical insights at an online workshop.
參與權限 Participation & Access
如何成為評測者？ How to become an evaluator?
所有註冊用戶預設皆具備評測者（學生）身份。登入後直接進入「控制台 (Dashboard)」並點擊「開始評測」即可進入競技場開始投票。
All registered users are evaluators (students) by default. After logging in, go to the "Dashboard" and click "Start Reviewing" to enter the Arena and begin voting.
評測次數有限制嗎？ Is there a limit to how many times I can vote?
沒有嚴格限制，但為了確保排行榜的公正性，系統會自動偵測異常投票行為。我們建議您在完整觀看並比較影片後再投下神聖的一票。
There is no strict limit, but the system automatically detects abnormal voting behavior to ensure fairness. We recommend voting only after fully watching and comparing the videos.
評測規則與標準 Review Rules & Criteria
具體的評分標準是什麼？ What are the specific scoring criteria?
請根據以下四個面向進行綜合判斷：
1. 內容正確性：資訊是否準確？是否有錯誤知識或幻覺？
2. 教學邏輯：解說是否由淺入深？結構是否清晰？
3. 適性化程度：是否符合該主題目標學生的認知程度？
4. 吸引力與投入感：敘事是否生動？視覺與聽覺是否同步且具吸引力？
Please judge based on these four aspects:
1. Accuracy: Is the information correct? Are there errors or hallucinations?
2. Pedagogical Logic: Does it explain from simple to complex? Is the structure clear?
3. Adaptation: Does it match the target student's cognitive level?
4. Engagement: Is the narrative vivid? Are the visuals and audio synchronized and attractive?
如果兩個影片都很優秀，或者都很糟糕怎麼辦？ What if both videos are excellent, or both are terrible?
競技場的設計是為了強迫選出「相對更好」的一個。即使兩者差異細微，也請嘗試根據上述標準找出表現較佳的影片。如果真的完全無法區分，系統亦提供「平手 (Tie)」的選項。
The Arena is designed to pick the "relatively better" option. Even if the difference is small, try to identify the superior one. If they are truly identical in quality, a "Tie" option is available.
如何參加 How to participate
如何成為參賽者？ How to become a contestant?
註冊帳號後，預設並未開通參賽權限。請前往「設定」頁面 (https://teaching.monster/app/settings)，勾選並同意「參賽者協議書」後，即可開通參賽者身份。
After registration, contestant permissions are not enabled by default. Please go to the "Settings" page (https://teaching.monster/app/settings), check and agree to the "Contestant Agreement" to activate your status.
有範例程式可以參考嗎？ Are there any sample programs I can refer to?
該 GitHub 儲存庫為官方團隊開發的一個基線模型管線，能順利運行從接收請求到生成影片回傳的每一個環節，若有需要歡迎參考。
This GitHub repository is a baseline model pipeline developed by the official team. It can run smoothly through every step from receiving requests to generating and transmitting the movie back. Feel free to refer to it if needed.
操作指南 Operation Guide
參賽者平台如何操作？ How to operate the contestant platform?
1. 新增模型：在「參賽者中心」選擇競賽，點擊「新增模型」並填寫模型名稱與您的 API Endpoint。
2. API 測試：填寫後建議利用平台的「測試功能」確保 API 溝通正常。
3. 生成影片：點擊「生成控制面板」，選擇想要測試的主題並點擊生成，完成後可進行預覽。
4. 公開模型：確認影片品質符合預期後，在模型頁面點擊「公開模型」，即可讓評測者開始對您的模型進行評測。
1. Add Model: Select a competition in the "Contestant Center", click "Add Model", and fill in the model name and your API Endpoint.
2. Test API: Use the platform's "Test" feature to ensure your API is communicating correctly.
3. Generate Video: Open the "Generation Control Panel", select a theme, and click generate. You can preview the video once finished.
4. Publish Model: Once confirmed, click "Publish Model" on the model page to allow reviewers to start voting on your videos.
提交規範 Submission Rules
上傳次數有限制嗎？ What are the submission limits?
為確保公平與評測資源分配，在競賽期間，會限制每個主題的生成請求間隔，預設為 5 分鐘，但仍可能隨平台負擔隨時調整。
To ensure fairness and the allocation of evaluation resources, the interval between generation requests for each topic will be limited during the competition, with a default of 5 minutes, but this may be adjusted at any time depending on the platform's workload.
我可以使用哪些素材？ What data can I use?
您可以使用網路上的第三方素材，但必須遵守以下規範：1. 清楚標示來源。 2. 嚴禁捏造不存在的數據（零幻覺）。 3. 優先使用 CC 授權或 AI 原創素材。
You can use third-party materials but must: 1. Clearly attribute sources. 2. Strictly forbid fabricating data (Zero Hallucination). 3. Prioritize CC-licensed or AI-original materials.
系統運算與媒體規格有什麼要求？ What are the technical requirements?
系統須在 30 分鐘內回傳影片。視訊格式需為 MP4，解析度至少 720p，音訊取樣率不低於 16 kHz。
The system must return the video within 30 minutes. Videos must be in MP4 format, at least 720p resolution, with an audio sampling rate not lower than 16 kHz.
一個隊伍可以提交多個模型嗎？ Can a team submit multiple models?
初賽期間每個隊伍只能建立一個模型(熱身賽可建立三個)，且同一個模型可以更新後重新生成影片。但初賽截止（5/15 00:00:00 AoE）後不可再重新生成影片。
During the preliminary round, each team can only create one model (three models are allowed during the warm-up round), and the same model can be updated and the video regenerated. However, the video cannot be regenerated after the preliminary round ends (May 15, 00:00:00 AoE).
排名與驗證 Ranking & Verification
最終排名是如何決定的？ How will final rankings be determined?
初賽將先依照 AI 學生評分選出最多 10 名，進行真人評測，最終選出前 3 名進入決賽，由頂尖專家評審團觀看影片後進行綜合排名，決定最終獲勝者。
The preliminary round will select up to 10 students based on AI student ratings for a real-person evaluation. The top 3 will then advance to the final round, where a panel of top experts will watch the videos and rank the teams to determine the final winner.
獲獎隊伍需要提供什麼驗證？ What verification is required for winners?
獲獎候選隊伍須於賽後提交完整的系統程式碼映像檔 (如 Docker)，以供主辦方在隔離環境中進行全自動化流程的複驗，確保無人工介入。
Winning candidate teams must submit a complete system image (e.g., Docker) for the organizers to verify the full automation process in an isolated environment, ensuring no human intervention.