Machine Learning 2026 Spring

23

詳細已公告於ntucool以及下方課程內容和作業區

課程加簽碼已全數寄出，再請同學查詢信箱

Date	Topic	Class Material	Extra Material
3/6	AI Agent - 1	機器學習2026規則說明 / ppt / pdf 解剖小龍蝦 — 以 OpenClaw 為例介紹 AI Agent 的運作原理 / ppt / pdf	【生成式人工智慧與機器學習導論2025】第１講：一堂課搞懂生成式人工智慧的原理
3/13	AI Agent - 2	Context Engineering 基本概念解說 / AI Agent 之間可以有什麼樣的互動/ AI Agent 對於工作帶來的衝擊 - 以學術研究為例/ ppt / pdf
3/20	深入模型內部架構：如何加快模型推論速度	加快語言模型生成速度 (1/2)：Flash Attention 加快語言模型生成速度 (2/2)：KV Cache/ ppt / pdf
3/27	深入模型內部架構：模型如處理超長輸入	Positional Embedding ppt / pdf
4/10	如何教育模型 - 1
4/24	如何教育模型 - 2
5/8	模型的自我成長 - 1
5/15	Appier Research 團隊演講
5/22	模型的自我成長 - 2
5/29	陳暐教授演講
6/05	生成策略 - flow matching

The information here is tentative and subject to change. Please read the requirement of each homework before deadline.

You should NOT plagiarize, if you use any other resource, you should cite it in the reference.
You should NOT modify your prediction files manually.
Do NOT share codes or prediction files with any living creatures.
Your final grade x 0.9 and get a score 0 for that homework if you violate any of the above rules first time (within a semester)
Your will get F for the final grade if you violate any of the above rules multiple times (within a semester).
Prof. Lee & TAs preserve the rights to change the rules & grades.

#	Date	Topic	Code	Platform	Deadline (UTC+8)	TA
x	3/6	Colab and Kaggle Tutorial		N/A	N/A	N/A
x	3/6	PyTorch Tutorial	N/A	N/A	N/A	N/A
x	3/6	JudgeBoi Guide	N/A	N/A	N/A	N/A
Bonus	3/13	Bonus Competition	N/A	N/A	05/15/2026 23:59	許筠曼
HW 1	3/6	LLM Malicious Instruction Defense			03/26/2026 23:59	董家愷、陳思齊、許筠曼
HW 2	3/13	AI Agent as an AI Engineer			04/02/2026 23:59	林禹融劉建蘴楊樂霖
HW 3	3/20	LLM Fast Inference		NTUCOOL	04/09/2026 23:59	馮柏翰, 吳岳霖, 蘇炳揚
HW 4	3/27	Training Transformer			04/16/2026 23:59	劉建蘴、馮柏翰、陳品睿
HW 5	4/10	Finetuning without Forgetting
HW 6	4/24	Model Editing
HW 7	5/8	Model Merging
HW 8	5/15	Test-Time Scaling
HW 9	5/22	Flow Matching
HW 10	5/29	Spoken Language Model

* 請同學來信務必依照正確寄信格式，並請善用NTU Cool討論區