機器學習
Machine Learning
| 節 | 週一 |
|---|---|
6 14:20–15:10 | 機器學習 3 節連堂 |
7 15:30–16:20 | |
8 16:30–17:20 |
* 根據陽明交大上課時間表所列
本課程的主要講授內容是機器學習的基礎理論與最新研究課題,特別著重在實際應用,因此將與來自聯發科技(MediaTeK)的業師在機器學習實務應用案例上共同授課。目標是使學生在修習本課程之後,建立機器學習的學理知識,同時具備自行建構機器學習系統以解決真實世界問題的能力,為以後從事相關領域的研究與開發工作奠定堅實的基礎。
無絕對必須之先修科目或先備能力的要求,但建議具備基本機率、線性代數知識者為佳。本課程之課程作業原則上將使用Python程式語言來進行。
因應疫情,本課程暫以線上方式授課(Google Meet 會議): 連結:https://meet.google.com/yvq-jnfx-vco
(subject to change) Assignments (40%) - 2-3 times, in-class or from the course website - If “invited talks” are arranged: Class Participation (10%) Midterm Oral Presentation (20%) - In-class presentation of the proposed topic or a target challenge for final project, with a related paper published in 2021-2022 from the given list of journals/conferences Final project (40%) - A live demonstration of self-developed machine learning applications or challenge solutions, along with an introductory slide or poster
| 週次 | 主題 |
|---|---|
| 第 1 週 | - Overview of Machine Learning |
| 第 2 週 | - Machine Learning Basics - Deep Feedforward Network |
| 第 3 週 | 放假 (和平紀念日) |
| 第 4 週 | - Regularization for Deep Learning - Optimization for Training Deep Models |
| 第 5 週 | - Self-Attention and Transformers |
| 第 6 週 | - Self-Supervised Learning |
| 第 7 週 | - Model Compression and Acceleration for Deep Neural Networks (I) |
| 第 8 週 | 放假 (民族掃墓節) |
| 第 9 週 | - Model Compression and Acceleration for Deep Neural Networks (II) |
| 第 10 週 | Midterm Presentation |
| 第 11 週 | Midterm Presentation |
| 第 12 週 | ML in Industry Practice (MediaTek) |
| 第 13 週 | ML in Industry Practice (MediaTek) |
| 第 14 週 | ML in Industry Practice (MediaTek) |
| 第 15 週 | ML in Industry Practice (MediaTek) |
| 第 16 週 | Final Project |
以教師自編講義與收集而得之公開學術資源為主,輔以下列參考書籍: Ian Goodfellow, Yoshua Bengio, Aaron Courville, “Deep Learning”, MIT 2016.