Home
Explore Uedu
Student Console
Register as Member/Login
研究知情同意中心
問卷中心
Teacher Console
Course Setup
支援與訊息
Uptime Data

UeduGPTs

--

Jupyters

0

Local AI

--

CISOSE26 本地 AI UG26
臺北 AQI 33 · 臺中 AQI 22 · 臺南 AQI 27 · 高雄 AQI 31

AI 回覆桌面通知

AI 助教回覆完成時顯示桌面通知

聊天訊息通知

同學在討論區發送訊息時通知

Sound notification

Play an alert sound whenever there is a new notification

Empowering Higher-Order Thinking[1][2][3]

Guiding students from passive learning to active critical thinking through AI-powered Socratic dialogue

[1] Evidence source:ACM Learning @ Scale 2026 full paper (Best Paper nominee)—an analysis of cognitive engagement across hundreds of university courses
[2] Evidence source:Computers & Education: Artificial Intelligence journal article (accepted, in press)—an analysis of cognitive-engagement patterns in students' AI conversations across disciplines
[3] Evidence source:Frontiers in Psychology journal article (published 2026)—effects of a dual-role AI lecturer/teaching assistant on learner autonomy, self-regulated learning, and intrinsic motivation in university EFL

Try without signing in: chat with the AI running on our campus GPUs ↓

Need an AI assistant for your course? How to apply

Latest updates
賀!Uedu「對話即學習(Chat as Learning)」研究獲國際期刊《Computers and Education: Artificial Intelligence》接受 台灣四百年上線:檔案式台灣史與轉型正義教材,附 3D 歷史地標島圖與 AI 陪讀 賀!Uedu「AI 助教規模化」入圍 ACM L@S 2026 最佳論文:全會議四篇入圍之一 賀!「生理感知語言模型(PALM)」獲國科會新進人員研究計畫補助 優學院走進研究現場:115 年度教學實踐研究與國科會大專生計畫採用 Uedu優學院 中央大學攜手 Garmin Health:讓 AI 與穿戴裝置成為「學界共享的研究基礎設施」 Uedu | Garmin Health:消費級穿戴裝置在教育與研究的應用研討會 多受試者生理資料即時擷取與事件對照:Educational Omics 的神經生理向度推進
More →
福爾摩沙
特別企劃

台灣四百年

台灣史與轉型正義・檔案式史實教材|對齊 108 課綱台灣史

四百年,是一連串外來政權更迭的長度;但這座島嶼的故事,遠比這更早。從原住民與史前為序章,走過荷西、明鄭、清領、日治,直到戰後的威權與民主化,以及面對過去的轉型正義。附 3D 歷史地標島圖、四百年大時間軸與 AI 陪讀,所有敘述力求客觀、附權威來源。

走進台灣四百年 3D 歷史地標島圖
動手做科學
特別企劃

Uedu Science 科學實驗室

互動式科學模擬・物理/化學/生物/地球科學|瀏覽器即開即用

九十四個互動式科學實驗,在瀏覽器裡直接動手做:調整參數、觀察現象、記錄並匯出數據。從單擺週期、拋體運動到波的干涉,涵蓋四大學科,全部即時運算、免安裝;登入後還有 AI 實驗助教陪你邊做邊問。

走進科學實驗室
Chat as Learning
Ask Well, Think Deeply
Computers and Education: Artificial Intelligence
Computers and Education:
Artificial Intelligence
IF 23.4 CiteScore 43.3 #2 worldwide in Education Open Access
2026 · Accepted, in press
Uedu’s second international journal article

Chat as Learning: Making Cognitive Engagement in AI Conversations Visible

Chat as Learning: Student–AI Conversations as Discipline-Associated Cognitive Engagement Patterns
C.-K. Chang* and K.-H. Li·Computers and Education: Artificial Intelligence, 2026 (accepted, in press)

Research vision: Uedu proposes the “Chat as Learning” measurement paradigm—every prompt a student poses to an AI teaching assistant is an observable externalisation of learning cognition in progress. The motto “Ask Well, Think Deeply” draws on “inquire carefully, reflect thoroughly” from the Doctrine of the Mean and the art of good questioning in the Record of Learning (Xueji); we use Bloom’s Taxonomy to analyse the cognitive engagement patterns in students’ prompts—not to grade performance, but to make thinking processes visible. The paper, by Assistant Professor Chia-Kai Chang and Mr. Kuei-Hao Li, appears in Computers and Education: Artificial Intelligence—the journal ranked #2 worldwide in Education (Impact Factor 23.4), dedicated to research on AI applications in education.

Analysis across four universities, two semesters, 116 courses and over 60,000 student messages shows that about 62% of prompts reach higher-order cognition; moreover, the same student shifts cognitive patterns when the discipline changes—Apply is most prominent in STEM courses (20.8%), Understand dominates in Language courses (31.7%), and Create stands out in Social Science courses (33.8%). The variance in higher-order engagement comes mainly from the course discipline rather than the individual student: cognitive engagement is a discipline-contingent pattern, not a fixed personal interaction style.

See the research →
*Metrics index the cognitive level of students’ prompts (Bloom), not learning outcomes; the within-person contrast is directionally consistent across two independent semester samples; Humanities courses, with a small sample, are reported descriptively in the paper. *Corresponding author.
Frontiers in Psychology
SSCI Q1 in Psychology Peer-reviewed
2026 · doi
Uedu’s first international journal article

Agentic-AI-Empowered Self-Directed University English Learning

Agentic AI as a Dual-Role Lecturer and Teaching Assistant: Effects on Learner Autonomy, Self-Regulated Learning, and Intrinsic Motivation in University-Level EFL Education
K.-H. Li and C.-K. Chang*·Frontiers in Psychology, vol. 17, Article 1847607, 2026

Background and course context: The study's first author is Mr. K.-H. Li, co-founder of Uedu, and it builds on empirical research conducted in his university English course—deeply integrating the dual-role Agentic AI learning environment built on the Uedu platform (AI Lecturer + AI Teaching Assistant) into the course design to examine its effects on learner autonomy, self-regulated learning and intrinsic motivation. Mr. Li's university English course was also selected as a model course; the study thus earns dual recognition—academic (the international SSCI journal) and pedagogical (a model course).

A 16-week controlled study (120 undergraduates; 60 experimental / 60 control) found that when AI guided learners through Socratic questioning and gradually faded its scaffolding as competence grew, students’ gains in learner autonomy, self-regulated learning and intrinsic motivation were significantly greater than the control group (large effect sizes); most students’ questions to the AI also shifted gradually from “seeking answers” toward “seeking elaboration and metacognition.”

See the research →
*A single controlled study (quasi-experimental, n=120); the between-group effect reflects both ‘AI access’ and ‘task format,’ indexing effects ‘with technology,’ and whether gains persist after AI scaffolding is withdrawn awaits longitudinal verification. *Corresponding author.

Chat with the AI on campus GPUs

Uedu’s self-hosted local model: conversations never leave campus, optimised for Traditional Chinese, with system load fully transparent. Try it without signing in.

Local AI Chat Taiwan-8B
Hi! I’m a local model running on campus GPUs—no commercial cloud involved. Ask me anything—Traditional Chinese is my strength.
🌐 Web search is on: your query keywords are sent to an external search engine (via an anonymising campus proxy); the conversation itself is still processed only on campus.

Research use notice: to improve model quality and support AI-literacy research, your conversations and system performance metrics are stored; when not signed in they are recorded anonymously, linked to no identity, and used only for aggregate analysis. Please do not enter personal or sensitive information. Learn about this model and the case for taking charge of AI →

Live system status
Service status
Generating--
Queued--
KV cache usage--%

This model runs on a single RTX 4090 GPU on campus—the real load of a one-card system, shown transparently. You may need to queue at peak times.

Teaching as usual, thinking deeper

AI Socratic dialogue centers on questioning, helping students naturally build critical thinking within your existing course

Pre-class preview

AI asks, understanding begins

  • Upload materials to auto-generate questions
  • Students think actively, not receive passively
  • Easier to get up to speed in class
Focused in class

Your teaching, with AI alongside

  • Keep your slides, keep your handouts
  • Students who preview interact better
  • More efficient class time
After-class review

AI asks, reflection deepens

  • Follow-up questions instead of answers, deepening reflection
  • Track each student's thinking trajectory
  • Automatically flag students who need attention
RESEARCH POWERED BY UEDU

Independent research teams choose Uedu

The strongest evidence for the platform is not our own claim—it is researchers across institutions independently designing and running studies with Uedu as their tool or research site, and publishing in international peer-reviewed journals.

JSLW · SSCI 2026 · 期刊論文 How the Uedu team is credited:具名誌謝

純提問式 AI 蘇格拉底對話作為第二語言論說文寫作的對話式回饋:一項準實驗研究

Question-only AI Socratic dialogue as dialogic feedback in L2 argumentative writing: A quasi-experimental study
L.-J. Wang·國立中央大學語言中心·Journal of Second Language Writing

實驗組在證據運用(η²ₚ = .255)與反駁整合(η²ₚ = .087)顯著優於對照組;立場(claims)在兩點量表下未見顯著差異。

This study was designed, conducted and published by an independent research team.作者於誌謝具名感謝張家凱博士提供平台與技術支援;平台於方法一節具名為 Uedu。

Talent for the AI Era

Not just a few technical experts, but cross-disciplinary AI literacy

The AI era needs more than a handful of top technical experts (depth); it needs an entire workforce that can collaborate with AI (breadth)—marketing, management, law, design: every role now calls for AI literacy. Uedu focuses on exactly this breadth: equipping students from every discipline to work alongside AI.

DEPTH
Specialized · Few · Deep
  • Specialists in AI and technology fields
  • Deep expertise, industry-ready
  • Cultivated by specialized colleges
×
BREADTH
Cross-domain · Everyone · Broad
  • Cross-disciplinary AI literacy (business, humanities, design)
  • General education courses × the Uedu platform
  • Able to use it, judge it, and say no
True “new AI literacy” = depth × breadth
We need both a few elite technical specialists and AI literacy across every discipline.
NEW · 2026 Focus

AI Literacy — use AI without being used by it

"Using AI" is very different from "having AI literacy"—just as knowing how to drive isn't the same as driving literacy. Uedu doesn't just let you use AI; we design it so the more you use it, the more you think, question AI, and judge for yourself.

This page integrates six international and local frameworks—UNESCO, OECD, the EU AI Act, AAAI, IEEE, and Taiwan's Ministry of Education—to explain what AI literacy is, why everyone must learn it after 2026, and how Uedu helps you practice, while documenting global moves such as Purdue University making AI literacy a graduation requirement in Dec 2025.

Open the full AI literacy guide Academic version
Ability 1
Understand AI
What AI is and what it can do
Ability 2
Use responsibly
The outcome is your responsibility
Ability 3
Protect data
What you shouldn't hand to AI
Ability 4
Question AI answers
Verify before you trust
Ability 5
Learn deeper with AI
Not just for speed
Ability 6
Create and refuse
Able to use it, brave enough to say no

The same skill as media literacy—just the AI version.

For teachers: how assessment changes in the AI era

AI has cracked summative assessment;
formative assessment fills the blind spot

Reports students write with AI increasingly look like real learning—final exams and written reports are losing their discriminating power. You can't tell which the student understood and which the AI wrote.

Formative assessment looks not at the final "output" but at the "thinking process"—where students get stuck, whether they ask follow-ups, whether they can critique AI answers. Process evidence doesn't lie.

Uedu's "Knowledge Trace"
lets your classroom accumulate formative evidence automatically

  • Automatically records every turn of dialogue, every follow-up, every pivot between students and AI
  • Auto-tags Bloom's cognitive levels, revealing students' thinking trajectory from remembering to creating
  • No extra tools needed—students' interaction with AI generates the trace naturally
  • Complements traditional summative assessments like final exams and reports—coexisting rather than replacing
Academic positioning: Uedu's knowledge trace is the process evidence that formative assessment needs; the teacher remains the assessor—the platform only accumulates data and visualizes the trace.
Agentic AI Learner Autonomy | Learner Autonomy

You lead, AI guides

Let Agentic AI guide thinking instead of moving ahead for the learner
“Learner autonomy” means learners can set their own goals, choose their methods, monitor their progress, and evaluate their outcomes—rather than passively waiting to be fed answers. Yet the more capable an AI is, the more easily it “over-helps”—writing the answers, reports, and code in full, so the learner loses the chance to think for themselves.

So Uedu does not build a “you ask, it answers” chatbot; instead it uses Agentic AI(agent-based AI)——an AI learning agent that proactively plans steps, asks the learner questions, calls tools to find information, and gradually adjusts how much help it gives based on the learner’s performance. It does not just answer questions—it “guides you to find the answer yourself.” You lead; AI only guides.
Uedu’s dual-role Agentic AI teaching assistant turns the spirit of “you lead, AI guides” into four actionable design principles:
Theory
Autonomy Support
SDT
Encourages self-determination, reflection, and learner-initiated responses
Theory
Zone of Proximal Development
ZPD
Provides adaptive scaffolding pitched just beyond independent ability
Core
Guiding by Questioning
Socratic
Guides elaboration and reasoning through questions, without giving answers directly
Core
Scaffolding Fading
Fading
Gradually reduces support as performance improves, returning responsibility to the learner

Uedu translates Self-Determination Theory and Vygotsky’s Zone of Proximal Development into the design of its Agentic AI teaching assistant: the agent uses Socratic questioning to guide learners to elaborate and reflect, and lets scaffolding fade gradually as ability grows. In a 16-week quasi-experiment, most learners’ questions to the AI shifted gradually from “seeking answers” toward “seeking elaboration and metacognition”—an observable trajectory of learning behavior.

16
Weeks (quasi-experiment)
90%
of learners shifted toward higher-order questions
120
university students (60 / 60 control)
*The metric is an “observable question-behavior trajectory”—a descriptive pattern, not learning outcomes or grades. The between-group effect also reflects differences in task format, and whether it persists after the AI scaffolding is withdrawn awaits verification by longitudinal research.
Empirical evidence for this approach has been published in 《Frontiers in Psychology》(an SSCI international peer-reviewed journal) · doi:10.3389/fpsyg.2026.1847607
See the research

What is Uedu?

Uedu is an EdTech platform built around generative AI, integrating teaching, learning and research. Through multidimensional Educational Omics data integration, it creates a personalized learning experience for every learner.

Cognomics Linguomics PhysioNeuromics Sociomics Environomics Ethicomics

Platform highlights

Uedu offers a complete set of teaching and research tools to help teachers improve teaching quality

Teacher tools
AI Quiz System

After you upload materials, AI auto-generates varied question types, with item-bank management, auto-grading and score analysis

Class Screen Recording

Record your screen and webcam, auto-saved and shareable to students for after-class replay

AI Knowledge Base

Upload course materials and the AI assistant answers students' questions based on your content

Garmin Wearable Integration

Collect HRV, sleep and stress data to explore how physiological state affects learning

Socratic dialogue: AI guides thinking through questions

Instead of giving answers, it uses layered questions to help students discover answers themselves

Ordinary AI chat

Direct-answer mode

Get answers fast
👤

What is photosynthesis?

🤖

Photosynthesis is the process by which plants use light energy to turn carbon dioxide and water into glucose and oxygen. The formula is: 6CO₂ + 6H₂O + light energy → C₆H₁₂O₆ + 6O₂

The student gets the answer fast but skips the thinking process
Uedu AI (Socratic mode)

Guide thinking, train the brain

✅ Build a knowledge trace
👤

What is photosynthesis?

🧠

Great question! Before I answer, let me ask you first:
What "materials" do you think a plant needs to grow?

👤

Um... water, sunlight, and air?

🧠

Excellent! You've named three key elements.
So, what role do you think "sunlight" plays in this process? Is it a material? Or...?

Building knowledge trace...

The student is thinking about the difference between "energy" and "material"

See your students' thinking process

Every round of AI dialogue is auto-tagged with Bloom's cognitive levels. From "Remember" to "Create," you can watch students' thinking climb from shallow understanding to higher-order reasoning.

Automatic analysis

AI tags each message's cognitive level in real time—no manual reading needed

Timely detection

Which students are stuck at remembering? Who can already analyze and evaluate? See it at a glance

Track growth

Compare cognitive distribution at the start and end of the term to quantify teaching impact

Bloom's cognitive level distribution
Example: General Physics, term 114-2
Create18%
+10%
Evaluate25%
+13%
Analyze32%
+14%
Apply55%
+20%
Understand78%
+18%
Remember92%
+7%
Start of term End of term +N% is the growth

Participating teachers

Search by name, school or field of expertise

Found 20 teachers

張家凱
Dr. Chia-Kai Chang

National Central University

Center for General Education | Assistant Professor

Founder Principal Investigator AI Digital learning Information Technology IoT
李奎皓
Kuei-Hao Li, PhD Candidate

National Tsing Hua University

International Interdisciplinary PhD Program | PhD Candidate

Co-Founder Computing Education Digital learning
陳鏗任
Dr. Ken-Zen Chen

National Yang Ming Chiao Tung University

Graduate Institute of Education | Associate Professor; Center for Open Educational Resources in Higher Education | Deputy Director

Education Digital learning
鍾曉芳
Dr. Siaw-Fong Chung

National Chengchi University

Department of English | Professor

Linguistics
鄭芳祥
Dr. Fang-Hsiang Cheng

National Central University

Department of Chinese Literature | Associate Professor

Humanities
涂藍云
Dr. Lan-Yun Tu

National Central University

Department of Chinese Literature | Adjunct Assistant Professor

Humanities
李明懿
Dr. Ming-Yi Li

National Central University

Language Center | Assistant Professor

Language education Linguistics
王立仁
Dr. Wang Li Ren

National Central University

Language Center | Assistant Professor

Language education Digital learning
江道一
Tao-Yi Chiang, PhD Candidate

National Central University

Department of Chinese Literature | Adjunct Lecturer

Humanities
朱珊瑩
Dr. Chu Shan-Ying

Chung Yuan Christian University

Department of Business Administration | Associate Professor

Management
楊奕農
Dr. Yi-Nung Yang

Chung Yuan Christian University

Department of International Business | Associate Professor

Business
許瑛玿
Dr. Ying-Shao Hsu

Ming Chuan University

Department of Counseling and Industrial-Organizational Psychology | Associate Professor

Education Psychology
黃淑玲
Dr. Shu-Ling Huang

Kaohsiung Medical University

Center for General Education / Center for Humanities and Arts Education | Associate Professor

Education Psychology
張靜宜
Dr. Ching-Yi Chang

Taipei Medical University

Department of Nursing | Associate Professor

Nursing
蔡懿玲
Dr. Yi-Ling Tsai

National Taipei University of Nursing and Health Sciences

Department of Health Care Management | Assistant Professor

Health Nursing
戴秀珍
Dr. Hsiu-Jane Tame

Chang Gung University of Science and Technology

Department of Nursing | Assistant Professor

Nursing
張靖卿
Dr. Ching-Ching Chang

National Changhua University of Education

Department of Special Education | Associate Professor

Special Education
賴建智
Dr. Chien-Chih Lai

National Dong Hwa University

Department of Physics | Professor

Physics

What teachers say

Real experiences from across disciplines

Frequently Asked Questions

Questions about use, features and research—all answered here

Go to FAQ

Can't find an answer? Contact us

Educational Omics

Uedu platform uses Educational Omics as its theoretical framework, integrating learning data across six dimensions—cognition, language, physiology/neuroscience, society, environment and ethics—to build a Trusted Educational Data Lake.

Cognomics PhysioNeuromics Sociomics Environomics Linguomics Ethicomics
Explore Uedu Labs
For researchers
  • Download multimodal learning data
  • Wearable physiological data integration
  • AI dialogue trace analysis
  • Learning-environment sensor data

Ready to Get Started?

Want to add AI Socratic dialogue to your course? You're very welcome!

Direct Use

Want to add AI Socratic dialogue to your course and boost students' thinking? Free for teachers and students—you can experience all features.

  • Verify your school email once to gain recognized teacher status
  • Claim or create courses freely, no approval needed
  • Use all features for free
  • Start or stop anytime
Go to course setup
Free for teachers and students

This platform is supported by National Central University, with all features provided free of charge.

About data use: When you use the platform, Uedu collects the necessary interaction and learning-process data under defined guidelines for system optimization and teaching-feedback analysis for instructors (for example, generating a course knowledge trail). This data is used solely to improve service quality and is never used for individual assessment or grading. Only if you have signed a research informed-consent form (see the research ethics page) will we use the data, after de-identification, for academic research analysis. In addition, Uedu also supports the collection of exercise- and physiology-related data, which is likewise used only for system optimization; it is used for research purposes only after the user has signed an IRB-approved research informed-consent form.

Or learn more first

Principal Investigator:Chia-Kai Chang, Assistant Professor | [email protected]