深度學習與實務

Deep Learning and Practice

學期

106-2

學分

3 學分

當期課號

5962

永久課號

IOC5184

開課單位

資訊科學與工程研究所

授課教師

彭文孝、吳毅成、陳永昇

校區

光復

類別

選修

上課時間表

節	週二	週四
N 12:20–13:10		深度學習與實務 EDB27 3 節連堂
5 13:20–14:10
6 14:20–15:10
A 18:30–19:20	深度學習與實務 EDB27 3 節連堂
B 19:30–20:20
C 20:30–21:20

* 根據陽明交大上課時間表所列

概述

(1) To understand the maths of deep learning techniques (2) To familiarize with deep learning tools, such as Caffe, Tensor Flow, Torch, etc. (3) To understand the latest developments and applications of deep learning techniques (4) To develop practical working systems

先修科目

Linear Algebra, Probability Theory, Machine Learning (suggested) 2IJK-EC015 (for Labs) 4XEF-EC015 (for lectures) NOTE: (1) The first lecture will begin on Feb. 22 (4XEF-EC015). (2) If you want to enroll in this course, you need to be present on Feb. 22 (4XEF-EC015) and submit your enrollment form in person (if you have not yet been enrolled successfully). (3) Be advised that if we have more students taking this course than we could afford, your final enrollment will be subject to review by all the instructors. (4) More details will be announced during the first lecture. Make sure that you don't miss it.

評分方式

Computer assignments 40% Paper presentation 20% Final project and presentation 20% Final written exam 20%

課程大綱

Introduction
Machine Learning Basics
Deep Networks
Deep Networks
Deep Networks
Deep Networks
Deep Learning Research
Deep Learning Research
Deep Learning Research
Deep Learning Research
Deep Reinforcement Learning
Deep Reinforcement Learning
Deep Reinforcement Learning
Transfer Learning and Applications
Paper Study and Presentation

週次計畫

週次	主題
第 1 週	A. Introduction
第 2 週	B. Machine Learning Basics  Linear Algebra  Probability and Information Theory  Numerical Computation  Machine Learning Basics
第 3 週	C. Deep Networks  Deep Feedforward Networks  Convolutional Networks
第 4 週	 Convolutional Networks  Optimization for Training Deep Models
第 5 週	 Regularization for Deep Learning
第 6 週	 Recurrent and Recursive Nets
第 7 週	D. Deep Learning Research  Linear Factor Models  Autoencoders
第 8 週	 Autoencoders  Representation Learning
第 9 週	 Structured Probabilistic Models for Deep Learning  Approximate Inference
第 10 週	 Approximate Inference Deep Generative Models
第 11 週	E. Deep Reinforcement Learning Intro. to RL (MDP/POMDP + TD Learning)
第 12 週	F. Deep Reinforcement Learning Policy Gradient + DQN
第 13 週	G. Deep Reinforcement Learning DQN Applications: Atari, AlphaGo and Robotics.
第 14 週	H. Transfer Learning and Applications
第 15 週	Paper Study and Presentation
第 16 週	Paper Study and Presentation
第 17 週	Paper Study and Presentation
第 18 週	Final Project Presentation

教科書

1. I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning, 1st Ed., MIT Press, Dec. 2016 2. R. S. Sutton and A. G. Barto, Reinforcement Learning: An Introduction, Nov. 2017