Machine Learning

J.-S. Roger Jang, CSIE Dept., National Taiwan University

Math & Optimization
1. Gradient descent: (EMBA-week01)
2. Genetic algorithms:
3. Simulated annealing:
4. Downhill simplex search:
5. Random search
6. Tabu search
7. Jensen's inequality:
8. Linear combination of random variables:
Fundamentals of Machine Learning (ML)
1. Intro to AI: (EMBA-week01)
2. AI's dilemma: (EMBA-week01)
3. Intro to ML: (EMBA-week01)
4. Datasets for ML:
5. Supervised learning
  - K-nearest-neightbor classifiers: (EMBA-week01)
  - Linear classifiers: (EMBA-week01)
  - Naive Bayes classifiers:
  - Quadratic classifiers:
  - GMMC:
  - SLP:
  - MLP & DNN:
  - Backpropagation:
  - SVM
6. Unsupervised learning (Clustering)
  - K-mean clustering:
  - Fuzzy c-mean clustering:
  - Hierarchical clustering:
7. Density estimation
  - Maximum likelihood estimate (MLE):
  - Gaussian mixture models (GMM)
  - Hidden Markov models (HMM):
8. Dimensionality reduction
  - Feature selection:
  - Feature extraction
    - PCA:
    - LDA
9. Distance metrics and loss functions
  - Distance metrics:
  - Triplet loss: https://www.youtube.com/watch?v=d2XB5-tuCWU
10. Dynamic programming
  - Intro to dynamic programming (DP):
  - DP for longest common subsequence (LCS):
  - DP for edit distance (ED):
  - DP for matrix chain product (MCP):
  - Dynamic time warping (DTW):
11. Performance indices
  - ROC & DEC (old version):
  - Performance indices:
12. Performance evaluation
  - K-fold cross-validation:
13. Walkthrough
  - A walkthrough of machine learning for leaf identification:
Advaced Topics in Machine Learning
1. More models
  - CNN
    - Intro to CNN (convolutional neural networks)
    - Convolution in CNN
    - CNN applications (EMBA-week02)
  - Ensemble learning (including random forests and XGBoost):
  - RNN & LSTM:
  - GAN
2. Start your AI group
  - Advice for introduction of AI
    - Set up sensors, database, and data management teams
    - Pick easy tasks first
    - Prepare high-performance computers for AI-related training & modeling
  - 7 Steps to Design, Build, and Scale an AI Product by Allie Miller
  - 10 Steps to Adopting Artificial Intelligence in Your Business
Unstructured Data in Machine Learning
1. Chinese word segmentation: (EMBA-week02)
2. Document classification: (EMBA-week02)
Applications of Machine Learning
1. Manufacturing
  - AI for manufacturing (1h)
  - Wafer failure map recognition (1h)
  - Depth reconstruction from SEM images (1h)
  - Predictive maintenance (1h)
  - Timing assistant (ANSYS)
  - High performance approximate calculation for interactive design and protytyping (Fujitsu)
  - Application of GANs for 3D structural topology optimization
2. Audio processing
  - Intro to audio signals:
  - Basic audio features:
  - Pitch tracking in time domain:
  - Singing voice separation (a short version for children):
  - Source separation in music
  - Speaker verification/identification
3. Image recognition
4. FinTech
  - Candlestick chart (K線圖):
  - Intro to stock prediction:
  - Technical analysis (including MA):
  - Technical indicator of RSI:
  - DP for trading (when stock price is fully known in advance):
  - Sharpe ratio:
  - Portfolio optimization:
  - 金融擁抱 AI 的機會與挑戰:
  - AI 專案在金融場景的美麗與哀愁:
  - AI for FinTech (EMBA-week02)
  - Fintech examples
  - Precision sales
5. Chatbots
  - Chatbots (EMBA-week02)
6. Fuzzy sets and fuzzy logic
  - Fuzzy sets (a short version for children):
Summary and demos
1. SOP for machine learning (EMBA-week02)
2. Summary of ML (EMBA-week02)
3. Project summary at MIR Lab (EMBA-week02)
4. Demos at MIR Lab
Tutorials on specific topics