[CS Foundation] AIML - 1 - introduction

Transcript

1. Artificial Intelligence and Machine Learning Cheng-Kuan Ou X-Village
2. Outline • Introduction to AI & Machine Learning • Machine Learning Framework • Types of Machine Learning ◦ Supervised Learning ◦ Unsupervised Learning ◦ Reinforcement Learning • Machine Learning Workflow ◦ Data Preprocessing ◦ Train & Testing Data ◦ Evaluation 2
3. The development of Artificial Intelligence reference: https://blogs.nvidia.com.tw/2016/07/whats-difference-artificial-intelligence-machine-learning-deep-learning-ai/ 3
4. First progress (1956-1974) preference: https://www.inside.com.tw/2016/05/26/history-of-chat-bot-instant-messaging 4
5. Input layer hidden layer output layer 5 Second progress (1980-1987) • Artificial Neural Network • Expert System
6. AlexNet (2012) 8 layers 16.4% 16-19 layers VGGNet-16 (2014) 7.3% 22 layers GoogleNet-16 (2014) 6.7% 6 Third progress (2012 ~ )
7. Human 5.1% Top-5 error rate ResNet-152 (2015) 152 layers!! 3.7% 7 Third progress (2012 ~ )
9. What makes AI success in this era? 9
10. What makes AI success in this era? • Explosive increase in data 10
11. What makes AI success in this era? • Breakthrough in computing - GPU 11 GPU consists of more ALU than CPU in a chip. video: https://www.youtube.com/watch?v=-P28LKWTzrI
12. AI Application 12 video: https://www.ted.com/talks/joseph_redmon_how_a_computer_le arns_to_recognize_objects_instantly?language=zh-tw#t-395546 (Photo
13. Why do we need machine learning? 13 先試想以下問題: 1. 我們要怎麼判斷一封電子郵件是垃圾郵件還是重要郵件？ 2. 我們要推薦什麼樣的商品給我們的顧客？ 3. 我們可以用先前學到的資料結構、演算法處理這樣的問題嗎? - Weakness of hand-crafted rules - hard to consider all possibilities - Data-driven approach - Machine Learning
14. What is Machine Learning Learning It is a cat! 14 cat dog bird Write a program for learning.
15. Machine Learning 15 Learning ＝ Looking for a function from data • Image Recognition • Speech Recognition • Play Games
16. Framework Hypothesis Function Set Model 16 h 1 ( ) = “cat” h 1 ( ) = “dog” h 2 ( ) = “cat” h 2 ( ) = “bird”
17. Training Data Hypothesis Function Set Model Goodness of function Framework 17 h 1 ( ) = “cat” h 1 ( ) = “dog” h 2 ( ) = “cat” h 2 ( ) = “bird” f( ) = “cat” f( ) = “dog” f( ) = “bird” Target Function, f : X → Y
18. Training Data Hypothesis Function Set Model Goodness of function Framework 18 h 1 ( ) = “cat” h 1 ( ) = “dog” h 2 ( ) = “cat” h 2 ( ) = “bird” f( ) = “cat” f( ) = “dog” f( ) = “bird”
19. Training Data Hypothesis Function Set Model Goodness of function Framework 19 Pick the best function f( ) = “cat” f( ) = “dog” f( ) = “bird” Target Function Target Function, f : X → Y
20. Training Data Hypothesis Function Set Model Goodness of function Framework 20 Pick the best function f( ) = “cat” f( ) = “dog” f( ) = “bird” Testing Data h*( ) = “cat”
21. Training Data Hypothesis Function Set Model Goodness of function Framework 21 Pick the best function f( ) = “cat” f( ) = “dog” f( ) = “bird” Testing Data Training Testing h*( ) = “cat”
22. Hypothesis Function Set Machine Learning Step Model 3 1 2 Goodness of function Pick the best function 22
23. And it just like... when we want to put an elephant into a refrigerator 1 2 3 23 Hypothesis Function Set Model Goodness of function Pick the best function
24. Machine Learning Types of Machine Learning Reinforcement Unsupervised Supervised • Labeled data • Direct feeback • Predict outcome /future • Unlabeled data • No feeback • Find hidden structure • Markov Decision Process • Reward function • Learn policy 24
25. Classification Supervised Regression The class label is discrete. The class label is continuous. Label Supervised Learning 25 input 1 input 2 label input
26. Supervised Learning - Regression • Goal : Learning a function for a continuous output. 26 - Housing price prediction size of house estimated price ( scalar ) feature label
27. Supervised Learning - Regression 27 - Stock Price Prediction - PM2.5 Prediction
28. Supervised Learning - Classification 28 • Goal : Learning a function for a discrete output. ( categorical ) - Email Spam filtering (Binary Classification) Header keyword Spam ? { Yes , No } Header keyword feature class label
29. Supervised Learning - Classification 29 (slides by Cheng-Te Li) feature1 feature2 class1 class2
30. Supervised Learning - Classification 30 - Document classification - Medical diagnosis - Categorical : not ill, cold, flu - Weather - Categorical : sunny, cloud, rain, snow
31. Supervised Learning - Structured Learning 31 • Structured Data - beyond classification - Machine Translation - Speech Recognition 歡迎來修ＡＩ課程
32. Dimension Reduction Unsupervised Clustering Unsupervised Learning 32
33. Classifcation - Given a training set of labeled objects, learn a decision rule. 33
34. Clustering - Given a collection of (unlabeled) objects, find meaningful groups. 34
35. Unsupervised Learning - Clustering - Grouping 35
36. Unsupervised Learning - Clustering - Similarity 36
37. Unsupervised Learning - Clustering • A grouping of objects such that the objects in a group (cluster) are ◦ Similar to one another in a group ◦ Different from the objects in other groups 37 (slides by Cheng-Te Li) - K-means
38. 38 Unsupervised Learning - Clustering • Application
39. (slides by Cheng-Te Li) 39 Comparison
40. 40 Unsupervised Learning - Dimension Reduction
41. 41 Unsupervised Learning - Dimension Reduction (quoted from Hung-yi Lee’s slides) • A digit is 28 x 28 dimensions.
42. 42 Unsupervised Learning - Dimension Reduction X function Z High Dimension Low Dimension • Feature Selection (quoted from Hung-yi Lee’s slides) • Principal Component Analysis (PCA) z = Wx
43. Unsupervised Learning http://ttic.uchicago.edu/~ klivescu/MLSLP2016/ (slides of Ian Goodfellow) 43 - Generation
44. Reinforcement learning(強化學習) is another learning paradigm in machine learning. • Inspired from control theory and animal learning. • The learning agent will look around the environment and make a decision. 44 Reinforcement Learning Observation, Environment Feedback Decision
45. Reinforcement Learning - inspiration How do we teach animal to learn expected reaction? Give reward 45
46. Reinforcement Learning - inspiration When it do something wrong, just give it penalty. 46
47. In theory, if we can modeled a decision problem as a Markov Decision Process (MDP) and we can apply reinforcement learning to handle such problem. we need to define the following component in MDP: • State Space: How many possibilities in observation? • Action Space: What can I do with the environment? • Reward function: How good or bad about the decision? • Transition: Which state do I transfer to in the next time? 47 Reinforcement Learning - MDP
48. Reinforcement Learning - MDP 48 At each step t the agent: • Receives scalar reward R t • Receives observaiotn O t • Executes action A t The environment: • Receives action A t • Emits observation O t+1 • Emits scalar reward R t+1 Image credits: David Silver’s reinforcement Learning course in UCL, UK.
49. Reinforcement Learning - MDP A MDP can be recorded as the following sequence: 49 We often call this sequence as Trajectory or Episode, denoted by Observation/State, Reward Action
50. Reinforcement Learning - objective The objective of reinforcement learning is to learn the policy that can maximize expected total rewards in different MDPs. 50
51. Reinforcement Learning - algorithms development Currently, the reinforcement learning algorithms are under development and have been tested in different kinds of environments: 51 Video Games Board Games (The state space of 19x19 Go is 10171) Robotics
52. Reinforcement Learning - resources Course: • The best choice to engaged in deep reinforcement learning is CS294, UCB. • Another famous course is David Silver’s reinforcement learning in UCL. • The deep learning course from 吳尚鴻老師 in NTHU. • The reinforcement learning/deep learning course from 李宏毅老師 in NTU. You can find wonderful materials on Youtube. • We also have short courses about reinforcement learning in NCKU: ◦ https://netdbncku.github.io/dsai/2018/ 52
53. Reinforcement Learning - resources Currently, there are just few reading materials about reinforcement learning: • Reinforcement Learning: An Introduction ◦ The second edition will be published recently, and you can find the online draft. • Algorithms for Reinforcement Learning ◦ This book describes more details about reinforcement learning algorithms. • Top Conference Papers ◦ ICLR, NIPS, ICML, AAAI, ... 53 BTW, before you learn reinforcement learning, you had better to learn the basics of machine learning.
54. What is deep learning? 54
55. Deep Learning Deep learning just means using deep neural network as your model(function approximator) in machine learning. 55
56. Deep Learning Recently, the deep learning won big success in many domains and it can be applied in different kinds of learning paradigm to handle unstructured data. • Supervised learning: object detection ... • Unsupervised learning: generative adversarial network(GAN), auto-encoder • Reinforcement learning: robotics, power distribution, chatbot, NLP ... 56
57. Object detection 57 Image Credits: https://medium.com/@jonathan_hui/real-time-object-detection-with-yolo-yolov2-2 8b1b93e2088
58. https://make.girls.moe/#/ 58 GAN - 二次元妹子生成器
59. GAN - image-to-image translation 59 Quoted from Isola et al., “Image-to-Image Translation with Conditional Adversarial Networks”
60. Style Transfer 60 Qutoed from Gatys et al., “Image Style Transfer Using Convolutional Neural Networks” (CVPR2016)
61. Style Transfer 61 https://www.youtube.com/watch?v=Khuj4ASldmU
62. Deep Reinforcement Learning - BiCNet 62 https://www.youtube.co m/watch?v=kW2q15M NFug
63. Deep Reinforcement Learning - AlphaGO Zero 63 https://deepmind.com/blog/alphago-zero-learning-scratch/
64. Deep Reinforcement Learning - OpenAI Five Train the agents to play DOTA2 and they can cooperate with each others. (2018/8/6) 64 https://blog.openai.com/openai-five-benchmark-results/
65. Deep Learning We will cover the basic contents of deep learning in the 3rd lecture, but there are some prerequisite: • The concept of Matrix, vector multiplication. • The concept of partial derivative and chain rule (knowing derivative is OK.) 65
66. Supervised Learning Unsupervised Learning Reinforcement Learning Paradigm 66
67. Supervised Learning Regression Classification Unsupervised Learning Reinforcement Learning Structured Learning Paradigm Task 67
68. Supervised Learning Regression Classification Linear Model Structured Learning Non-linear Model Deep Learning decision tree, SVM, ... Unsupervised Learning Reinforcement Learning Paradigm Task Method 68
69. Supervised Learning Regression Classification Linear Model Structured Learning Non-linear Model Deep Learning decision tree, SVM, ... Unsupervised Learning Reinforcement Learning Paradigm Task Method 69
70. Hypothesis Function Set Machine Learning Step Model 3 1 2 Goodness of function Pick the best function 70
71. Hypothesis Function Set Goodness of function Training Data Model Testing Data Pick the best function Cost Function Cost Function : 71 Example:
72. Hypothesis Function Set Goodness of function Training Data Model Testing Data Pick the best function Cost Function 72
73. Hypothesis Function Set Goodness of function Training Data Model Testing Data Pick the best function Cost Function 73 Use optimization method to find optimal parameters of hypothesis!
74. Optimization Method • There are many optimization method to find optimal parameters of hypothesis. ◦ Gradient Descent/Ascent ◦ Least Square Method ◦ Linear programming/Quadratic Programming ◦ Newton’s Method ◦ Genetic Algorithms 74
75. Hypothesis Function Set Goodness of function Training Data Model Testing Data Pick the best function Cost Function 75 Gradient Descent
76. Gradient Descent https://zh.wikipedia.org/wiki/%E5%AF%BC%E6%95%B0 微分： 某一點的微分代表的是該點的斜率，也就是局部的變 化量。 76
77. Gradient Descent • Choose an initial value , where 0 denotes the number of iteration. 77
78. Gradient Descent • Choose an initial value . • Compute . 78
79. Gradient Descent • Choose an initial value . • Compute . learning rate : 79
80. Gradient Descent • Choose an initial value . • Compute . • Compute . learning rate : 80
81. Gradient Descent • Choose an initial value . • Compute . • Compute . learning rate : 81
82. Gradient Descent • Choose an initial value . • Compute . • Compute . learning rate : iteration... 82
83. Gradient Descent • Choose an initial value . • Compute . • Compute . Local Optimal learning rate : iteration... 83
84. Gradient Descent • Choose an initial value . • Compute . • Compute . Local Optimal Glogbal Optimal iteration... learning rate : 84
85. Machine Learning Workflow 85 https://www.linkedin.com/pulse/machine-learning-lets-learn-neeraj-sabharwal/
86. Get Data • Data analysis competition platform ◦ kaggle • Open data ◦ 政府資料開放平臺 ◦ 86
87. Data Pre-processing - Why should we pre-process data? "Garbage in, garbage out." 87
88. Data Pre-processing - Data type constraints: int, string, boolean - Accuracy: outlier, noise… - Completeness: missing value… - Feature Scaling 88
89. 89 Training & Test Data (slides by Cheng-Te Li)
90. 90 Training & Test Data All Data Training Data Test Data Training : Test 80% : 20% 75% : 25%
91. 91 Cross-Validation
92. Cross-Validation 92 (slides by Cheng-Te Li)
93. 93 Evaluation
94. = what model says true 94 Evaluation
95. Data Science in Python 95
96. 96 Scikit-Learn
97. 97 Thank you for listening.
98. 98 • Hung-yi Lee Course • Andrew Ng Course • Cheng-Te Li Course Reference
99. Reference 99