清華大學計科所
人工智慧程式設計
Artificial Intelligence Programming
ICMS National Tsing Hua University
劉晉良
Jinn-Liang Liu

2017 Fall Course

Lecture: PHYS物 504, 1:20pm, Mondays (2017.9.11 – 12.25)

Seminar: PHYS物 504, 2:45pm, Mondays (2017.9.11 – 12.25)

Grading: Coding Projects and Reports 100%

Reports: Send reports in ppt (short) and pdf (long) files to jinnliu@mail.nd.nthu.edu.tw on 10/15; 11/15; 12/25.

 

工業革命：動力1.0 (1760)，電力2.0 (1870)，數位3.0 (1945)，智力4.0 (2016 AlphaGo)

AI 聽說讀寫算行戲猜創醫研…, I/O AI, What AI, How AI. ADA.

 

Part I   Computer Programming (Browse and Use)

1.        Python Programming

2.        TensorFlow Programming (W1: Proj1: tf1.py, tf2.py, tf3.py)

 

Coding Notes

1. tf1.py: # Install Anaconda3-4.2.0 => Install PyCharm Community Edition 2017.1.4
# PyCharm => File => New Project => Location D:\AI\TF => Interpreter (Far Right Button) Click on Create Conda Env
# => Tools => Python Console => >>>  import pip => >>> pip.main(['install', 'tensorflow']) 
# See YouTube Link for more. # Put this file tf1.py in \TF. Run it. Done.
2. tf2.py: # How to use TensorBoard. # PyCharm => View => Tool Windows => Terminal => tensorboard --logdir="./graphs" --port 6006
# Google Chrome => http://localhost:6006/ => GRAPHS
3. tf3.py (Proj1: Google Code: mnist_softmax.py): # Google Chrome => http://localhost:6006/ => GRAPHS
4. mnist_1.0_softmax.py (Proj2: Google Code): # Error: matplotlib not installed
# PyCharm => File => Setting => Project: TF => Project Interpreter => + => matplotlib => Specify version => Install # Save zout_MNIST_1.0.png

 

Part II   Artificial Intelligence (Read and Work)

1.       A Simple Learning Model: Classification, Target, Hypothesis, Training Data, Learning Algorithm, Weights, Bias, Supervised and Unsupervised Learning

2.       Google Tutorial for ML Beginners: Image Recognition, MNIST, Softmax Regression (92%), Cross Entropy, Gradient Descent, Back Propagation, Computation Graph (Proj1: TensorFlow Code 1, Link)

3.       Tensorflow and Deep Learning I by Martin Gorner: Deep Learning Network, ReLU, Learning Rate (98%), Overfitting, Dropout (98.2%), Convolutional Neural Network (98.9, 99.3%)  (Proj1, 2, 3, 4: TensorFlow Code 1, 2, 3, 4, Download)

4.       Tensorflow and Deep Learning II by Martin Gorner (RNN1): Batch Normalization (99.5%), Data Whitening, Fully Connected Network, TensorFlow API, MNIST Record (99.7%), Recurrent Neural Network, Deep RNN, Long Short Term Memory, Gated Recurrent Unit, Language Model, Cloud Machine Learning (Proj5: TensorFlow Code 3)

 

Part III   AI Seminar

1.       2017.8.1: Udacity Open Source Self-Driving Car Project, Simulator, Auto1-Robocar-Lidar, Auto2-CNN-KITTI-TORCS, Auto3-CNN-OSM-GSV

2.       2017.8.8: D. Lee, Introduction to Humanoid Robotics

3.       2017.9.11: 陳冠霖, Introduction to Visual Common Sense for Autonomous Driving

4.       2017.9.25: Auto4-DeepQNet-Atari (DQN1), Auto5-DQN-TORCS (Patent, TFCode, DQN2, OpenAI Gym (Paper))

5.       2017.9.25: 劉冠漢, At First Glance for Direct Perception and Simulator TORCS in Autonomous Driving

 

More …

[1]* Google TensorFlow Frontiers, Non-Experts, ML APIs, Dev Summit 2017

[2]* OpenAI, OpenAI Universe

[3] TensorFlow Paper, Tensorflow Course

[4] Raúl Rojas (TOC), Yaser Abu-Mostafa

[5] Python Book, GPU Programming, GPU More

[6] Theano Paper, Theano Programming

[7]* KDD Cup Competition

[8]* Data Science Game

 

News

1.     2017.9.26: 科技部106年度第二梯次「博士創新之星計畫(LEAP)」即將於106年9月30日線上申請報名截止。科技部AI創新專案計畫推動辦公室剛獲悉有以下AI公司提供若干名額：－Deeplearning.ai (Andrew Ng 的新創公司) －Fortemedia (黃炎松博士創立專注語音處理技術) 為培育高階科技新興人才、引領人才連結國際並建立創新平台，惠請鼓勵合適人選踴躍報名參加。詳細資訊及報名方式請參考博士之星(官方網站)：https://leap.stpi.narl.org.tw/index.htm.

 

2017 Summer Course

A Short Scientific Programming Course for All Students

Prerequisites: Undergraduate Calculus and Programming

Time: 1 ~ 4pm, Tuesdays, July 4 ~ Aug. 8 (6 weeks), 2017

Place: 校本部綜三館203

Please register online 線上註冊 for this course.

Instructor: Jinn-Liang Liu 2017.5.21