Taiwan AI Open Courses Beyond NTU: Using TAICA's Course Lists to Find NTHU, NCCU, NCKU, and NTUT Classes

🌏 中文版

The NTU map covered Hung-yi Lee, Hsuan-Tien Lin, and Yun-Nung Chen. Outside NTU, Taiwan has several more AI courses taught in Mandarin with a full semester of recordings on YouTube. They are hard to find because the materials are scattered across instructors' GitHub repos, Google Sites, lab pages, and sometimes the website of a co-instructor at another university.

This post ties them together with one official source: the course list that TAICA (Taiwan AI College Alliance) publishes every semester. Ratings follow the A0–A3 scale from the Global AI/CS Course Map. A0 means only a catalog entry, A1 a syllabus, and A2 some substantive material or recordings. A3 means materials plus assignments are enough to form a coherent self-study path. The scale is this site's editorial judgment. It says nothing about credit or TA grading. NTU courses stay in the NTU post and are not repeated here. Everything below reflects checks made on September 30, 2026.

All courses in this post are taught in Mandarin unless noted otherwise.

What TAICA is: a list that tells you where each course streams

The "Origin of the Project" section on the TAICA homepage says the alliance was set up by the Ministry of Education. It pools teaching resources through cross-university AI programs, so universities with enough AI faculty can support schools that lack them. The alliance has run since academic year (AY) 113, which began in fall 2024, and the homepage says 55 institutions had joined by the second semester of AY113. The programs cover Applied AI Exploration, Industrial Applications, Natural Language Technology, and Computer Vision, and the site menu later added a Cybersecurity Technology program.

The model is built on "flagship courses" (主導課程). One university teaches the course and streams it live; other member schools open a matching local course for their students. The Info for Administrators page defines three authorization types:

• Closed authorization (mirror course): the flagship instructor handles all assessment for member-school students.
• Conditional authorization (satellite course): assessment follows the flagship instructor's design, while a co-instructor at the member school runs exams and grading.
• Open authorization (satellite course): the co-instructor can customize assessment.

For outside learners, the most useful pages are the semester course lists: Fall AY113, Spring AY113, Fall AY114, Spring AY114, and Fall AY115. Each entry lists the school, instructor, language, a syllabus PDF, and an "Online Class Link." Fall AY115 (fall 2026) has 10 flagship courses from six universities.

The "Online Class Link" field works as a filter. When it points to a YouTube channel, the stream and recordings are usually open to anyone. When it points to NTU COOL, Google Meet, or Webex, recordings usually stay with enrolled students. The field is only a starting point, though, and each course needs checking:

• Hung-Yu Kao's NLP at NTHU lists NTU COOL, yet its recordings are public on YouTube and linked from the GitHub course page.
• Yen-Lung Tsai's course at NCCU lists a Facebook live-stream group and separately names a YouTube channel as the recording archive.
• Kun-Ta Chuang's Generative AI Application Systems and Engineering at NCKU (Spring AY114) listed a YouTube stream for week 1 and NTU COOL for every other week. That week-1 video is now private.

At a glance: courses outside NTU worth self-studying

NTHU, Hung-Yu Kao's Natural Language Processing: the most complete course outside NTU

Kao's course page is the GitHub repo IKMLab/NTHU_Natural_Language_Processing. The front page is the 2026 edition, and earlier years sit in year folders. It is a graduate-level TAICA flagship course and has appeared on every fall list since Fall AY113. The Fall AY115 TAICA syllabus marks it as a mirror course, streamed live on Thursday mornings.

The latest complete term is Fall 2025. The schedule runs from word embeddings, language models, seq2seq and attention, Transformers, the BERT family, and decoding strategies to GPT-3, InstructGPT, RLHF, PEFT, and RAG. Each week links one or two YouTube recordings, 36 unique video links in total. The assignment index lists four assignments: Word Analogy, Arithmetic, Multi-output learning, and RAG. Each has an explainer video, a PDF brief, a report template, and a main.ipynb. TA sessions add example notebooks for PyTorch, Hugging Face, and LLM APIs.

That makes Fall 2025 an A3. What you cannot get is solutions, scores, and the full term-project spec. Fall 2026 is in progress with only W3 and HW1 posted so far, and its schedule ends with a new Reasoning/Agent unit.

This site's Reading NTHU Hung-Yu Kao Natural Language Processing series walks through Fall 2025 week by week, and its entry post compares grading and materials across both terms.

Something to do tonight: open Assignment 1 in the 2025 folder, watch the explainer video, and run the Word Analogy main.ipynb once in Colab.

NCCU, Yen-Lung Tsai's Generative AI: the fullest course page lives at Chang Gung

The course is Generative AI: Text and Image Synthesis Principles and Practice, taught by Yen-Lung Tsai of NCCU's Department of Mathematical Sciences. It is a TAICA flagship course. The Fall AY115 list marks it as a satellite course with a two-star difficulty, one of the most beginner-friendly AI courses on the list.

Search for this course and the most complete page you find is GenerativeAI2025. It sits on the CGU AICV Lab site of Chih-Yuan Yang in Chang Gung University's Department of Artificial Intelligence, and it is the page for Chang Gung's satellite class. The page states the offering school is NCCU and the instructor is Yen-Lung Tsai. Yang appears under co-instructor office hours, alongside two Chang Gung TAs. So the recordings and slides come from Tsai, while Yang supports and grades the Chang Gung section. That is TAICA's normal satellite arrangement, not two instructors co-teaching one course. The Chang Gung section also changed the grading: the average of 12 assignments counts for 100% and the final project for 0%, because seniors' grades were due early. NCCU's own grading follows the Fall AY115 syllabus.

The most complete term is Spring 2025 (1132):

• The 1132 playlist has 14 recordings, running from neural networks, GANs, Transformers, LLMs, chatbots, RAG, and AI agents to VAEs, diffusion, and ControlNet.
• The 14 slide PDFs are in Tsai's public Google Drive folder.
• The Chang Gung satellite page lists each week's assignment brief and rubric.
• Class demos live in yenlung/AI-Demo. The repo is shared across his courses and workshops and still changes after the term ends, so it is not a snapshot of 1132.

Fall 2025 (1141) was never gathered into a playlist. The live section of the Iveai – I've AI channel still holds that term's lectures 3 through 15. No public assignment page exists for that term, so it rates A2. Fall 2026 (1151) is in progress, and its playlist already holds the first four lectures.

This site's Reading NCCU Yen-Lung Tsai's Generative AI series follows 1132 lecture by lecture.

NCKU, Wei-Ta Chu's Introduction to AI: full recordings, half the materials

Chu's Introduction to Artificial Intelligence has been on the TAICA list since its first semester. The Fall AY115 syllabus gives a class size of 2,850. It streams on his YouTube channel and uses AIMA, 4th edition.

The URL is the catch. The course page path still says 2025f_AI, but the content is now Fall 2026, and the news line reads "Sep. 9, 2026 Website online." The Lectures page links slides only from Lecture 0 through Chapter 4; later chapters are commented out in the HTML. The assignments page shows only HW1, a final-project proposal. The syllabus lists five assignments, including programming work, and the other four specs are not public.

The recordings are in better shape. The 2025 playlist has 26 videos: one class recording per week plus a set of unit videos on attention, Transformers, CLIP, running an LLM locally, and RAG. The 2024 playlist has 15, and the 2024 course page likewise links only Ch0–4 slides. Overall it rates A2. It works well for an AIMA-style AI intro in Mandarin, but you will need to find your own exercises.

NTUT, Ping-Hsuan Han's Human-AI Interaction: AI as an interaction-design problem

This course takes a different angle from the others. It does not teach how to train models; it teaches how people interact with AI systems. The course site lists it as graduate-level, open to juniors and above, and taught in Mandarin. It streams on the XRLab channel of Han's lab.

The AY114-1 lecture playlist has 14 videos, and the channel has separate playlists for the midterm, final-project proposals, and final projects. The AY115-1 schedule lives in a public spreadsheet. The first half covers HCI basics, human-AI dialogue, human factors, and large multimodal models; the second half covers virtual humans, heuristic evaluation, and human-AI co-creation. Google Slides and recordings for the first three weeks are already linked, and the slides export straight to PDF.

The assignments page lists five prompts, such as "When AI-generated images fall short" and "When MCP is out of my control." Each is a short personal essay plus an online quiz, both run on NTU COOL, so outside readers can only take the prompts and write on their own. It rates A2.

NTHU, Min-Chun Hu's Robotic Navigation and Exploration: a robotics course found on the TAICA list

This course was not on the original shortlist; it turned up while scanning the Spring AY114 list. The list marks it as a Mandarin, graduate-level satellite course streaming on the NTHU RNE channel. The TAICA syllabus splits the content into three parts: SLAM, scene understanding with machine learning, and action control through path planning, navigation, and reinforcement learning. The reference texts include Probabilistic Robotics and Sutton & Barto.

The channel's live section holds 15 Spring 2026 recordings, Week 1 through Week 13. The curated RNE 2026 playlist stops at Week 5, so use the live section for the second half. The syllabus schedules a lab most weeks, but the only lab code online is in repos that students posted themselves. With no official lab specs or starter code, it rates A2.

NYCU: why it is not in the main table

NYCU has run TAICA flagship courses every semester since Spring AY113. The AI courses found for this post are all taught in English, and their recordings are not open enough:

• Deep Learning (Wen-Hsiao Peng, Yong-Sheng Chen, Ping-Chun Hsieh; Spring 2026): the Spring AY114 list points mainly to Google Meet and links a single YouTube video. That video is on a TA's personal channel and is unlisted, and the channel has no public videos. The only thing outsiders can rely on is the TAICA syllabus, so it rates A1.
• Physical AI (Yi-Ting Chen; Fall 2026): the playlist is public and currently has six videos covering Lectures 1–4, and the syllabus is public too. But the course is only four weeks in, and no course site or assignments turned up. It rates A2 (in progress) for now and is worth revisiting after the term.
• Introduction to Programming (C++) (Hung-Pin Wen): the list links a YouTube playlist, but this is a programming course rather than an AI course, and the playlist shows nine hidden videos.

Other TAICA courses: visible, but not this site's focus

Three reliable routes for outside learners

1. For NLP and LLM applications: NTHU Hung-Yu Kao's NLP, Fall 2025

The four assignments go from word vectors to RAG, each with a brief, template, and notebook. Its scope is close to Yun-Nung Chen's ADL at NTU, but more of its assignment material is public.

2. For beginners who want to build something first: NCCU Yen-Lung Tsai's Generative AI, 1132

Follow the weekly schedule on the Chang Gung satellite page: one recording and one assignment per week. It is shallower than Hung-yi Lee's courses, but most lectures come with a Colab assignment, and each one leaves you with a small working app.

3. For an AI survey or a different angle: NCKU Wei-Ta Chu, NTUT Ping-Hsuan Han, NTHU Min-Chun Hu

Chu's course gives you an AIMA-style tour in Mandarin, Hu's covers robotics and reinforcement learning, and Han's covers how users actually interact with AI. All three offer recordings and partial materials only, so you will need your own exercises.

The public courses outside NTU share one trait. TAICA needs thousands of students across schools to attend at once, so instructors stream to YouTube, and openness is a side effect. That is why recordings are complete while assignments and grading stay on NTU COOL. To keep up with new courses, open TAICA's latest course list at the start of each semester and read the "Online Class Link" column.

References

• Global AI/CS Course Map (series overview and A0–A3 definitions)
• NTU AI/ML Course Guide
• Harvard AI/ML Course Guide
• Reading NTHU Hung-Yu Kao Natural Language Processing: series entry
• Reading NCCU Yen-Lung Tsai's Generative AI: overview
• TAICA homepage (in Chinese, with English labels)
• TAICA: About Alliance Institutions (in Chinese)
• TAICA: Info for Administrators (authorization types) (bilingual)
• TAICA course list, Fall AY113 (in Chinese)
• TAICA course list, Spring AY113 (in Chinese)
• TAICA course list, Fall AY114 (bilingual)
• TAICA course list, Spring AY114 (bilingual)
• TAICA course list, Fall AY115 (bilingual)
• IKMLab/NTHU_Natural_Language_Processing (GitHub)
• NTHU NLP 2025 schedule
• NTHU NLP 2025 assignment index
• NTHU NLP Fall AY115 TAICA syllabus (in Chinese)
• Chang Gung satellite class: Generative AI 2025 (in Chinese)
• Yen-Lung Tsai, 1132 Generative AI playlist (in Mandarin)
• Yen-Lung Tsai, 1132 slide folder (in Chinese)
• yenlung/AI-Demo (GitHub)
• Iveai – I've AI live section (in Mandarin)
• NCCU Generative AI Fall AY115 syllabus (in Chinese)
• NCKU Wei-Ta Chu, Introduction to AI course page (currently Fall 2026)
• NCKU Wei-Ta Chu, Introduction to AI lectures
• NCKU Wei-Ta Chu, Introduction to AI 2024 lectures
• NCKU Introduction to AI Fall AY115 TAICA syllabus (in Chinese)
• Wei-Ta Chu YouTube channel (in Mandarin)
• Introduction to AI 2025 playlist (in Mandarin)
• Introduction to AI 2024 playlist (in Mandarin)
• NTUT Human-AI Interaction course site (in Chinese)
• NTUT Human-AI Interaction assignments page (in Chinese)
• XRLab NTUT YouTube channel (in Mandarin)
• AY114-1 Human-AI Interaction playlist (in Mandarin)
• NTHU Robotic Navigation and Exploration TAICA syllabus (in Chinese)
• NTHU RNE YouTube channel (in Mandarin)
• RNE 2026 class recordings playlist (in Mandarin)
• NYCU Deep Learning TAICA syllabus (in Chinese)
• NYCU Physical AI Fall 2026 playlist
• NYCU Physical AI TAICA syllabus (in Chinese)
• Tunghai AI Ethics stream channel (in Mandarin)
• Tunghai AI Ethics 1132 playlist (in Mandarin)
• NTHU ISA5810 Data Mining 2026 course page
• NTHU ISA5810 YouTube channel