UCL-AI-Society-ML-Tutorials

Season 1 (2020/2021)

Season 2 (21/22) Repo is right here -> https://github.com/UCLAIS/ML-Tutorials-Season-2

Contributors

• Danny Toeun Kim (Maintainer since 19/20)
• Jianqiao Mao (20/21)
• Kamen Brestnichki (20/21)
• Heejoon Koo (19/20)
• Maria Kapros (19/20)

Media

On our YouTube channel, you can watch our lectures and recordings of solution walk-through.
YouTube channel: UCL Artificial Intelligence Society

On our Discord channel, you can ask any questions you may have on our notebooks.
Discord channel invite link: UCL AI soc Discord

Check our Facebook events for the zoom link of the live session - AISoc.ucl

Tutorial notebooks & live lecture recordings

Release Date	Title	Solution video	Contents	Remarks
22th Oct 2020	Numerical Computation and Visualisation	Notebook 1 solution	Numpy, Pandas and Matplotlib	movie data NYC taxi data
5th Nov 2020	Supervised Learning	Notebook 2 solution	Regressions, Classification and Cross Validations	N.A.
19th Nov 2020	More on supervised learning and Intro to Unsupervised learning	Notebook 3 solution	SVM, PCA, K-Means	N.A.
3rd Dec 2020	Intro to Deep Learning	Notebook 4 solution	Perceptrons, Tensorflow, Keras and Pytorch	MNIST Digits from Kaggle

Pre-recorded lecture videos

Release Date	Title	Lecture video	Contents	Slides
5th Nov 2020	Introduction to Machine Learning	Lecture 01	Introduction, Data preprocessing, Mathematical Fundamentals, Linear Regression	Slide 01
19th Nov 2020	Building Blocks of ML	Lecture 02	Optimization, Linear algebra, Kernel Tricks	Slide 02
3rd Dec 2020	Supervised Learning (Advanced)	Lecture 03	KNN, SVM, Random Forest	Slide 03

Specific Contents

1. Machine Learning

1) Machine Learning Basics

1. Numpy, Pandas, Matplotlib

2. Exploratory Data Analysis (EDA)

3. Scikit-learn Basics
   a. Train test split
   b. Cross validation
   c. K-fold cross validation
   d. How to prepare your data before training

4. Model Evaluation (Can skip this)
   a. Accuracy, precision, recall
   b. ROC Curve and AUC
   c. F1-Score

---

2) Supervised Learning

5. Regression
   a. Linear Regression
   b. Multivariate Regression
   c. Ridge, Lasso, ElasticNet
   d. Boston house price prediction

6. Classification
   a. Logistic Regression
   b. Support Vector Machine
   i. Linear SVM
   ii. Kernel SVM
   c. Naïve Bayes

7. Decision Trees and Ensemble
   a. Binary Tree Decision Tree
   b. Hyper-parameter Tuning with GridSearchCV()
   c. Ensemble

---

3) Unsupervised Learning

8. Unsupervised Learning: Clustering
   a. K-Means Clustering

9. Primary Component Analysis (PCA) and Dimensionality Reduction

---

2. Deep Learning

4) Artifical Neural Networks

10. Perceptron
    a. SLP
    b. MLP

11. Tensorflow 2.0 + Keras Intro
    a. Now 2.0 has changed from beta to stable version so time to learn this
    b. Tensor Constant, Data type, LR with Tensorflow
    c. Keras Sequential, layers, model.compile()

12. Optimization (if time permits)
    a. Back Propagation
    a. Gradient Descent
    b. Stochastic Gradient Descent
    c. Adagrad
    d. Adadelta
    e. RMSprop
    f. Adam
    g. RAdam

13. Regularization Techniques
    a. Overfitting
    b. Data Augmentation (if time permits)
    c. Drop-out
    d. Batch Normalisation (if time permits)

14. CNN
    a. Data Augmentation for Computer Vision
    b. Convolution Layer
    c. Pooling
    d. Feature Map

15. Popular CNN Architectures
    a. AlexNet
    b. VGGNet
    c. GoogleNet
    d. ResNet
    e. Inception

16. Application of CNN - Style Transfer

17. Vanilla RNN

18. Long Short Term Memory (LSTM)

19. Bidirectional RNN

20. Tokenizing

21. Stopwords

22. Stemming

23. TF-IDF

24. Word Embedding

25. Word2Vec

26. A simple Chatbot

---

5) Reinforcement Learning

27. Elements of a RL problem and mathematical definitions

28. Morkov Decision Processes Formalism

29. Policy gradient algorithms
    a. REINFORCE
    b. Advantage Actor Critic (A2C)
    c. Asynchronous Advantage Actor Critic (A3C)

30. Q-learning
    a. SARSA
    b. DQN
    c. DQN Extensions

• Tutorial 1
  Notes
  solutions