Introduction to AI, AI Programming & Machine Learning

Outline

1. What is AI and Machine Learning

· Is machine learning difficult?

· What is artificial intelligence

· Difference between AI and machine learning

· Machine learning examples

2. Types of Machine Learning

· Three different types of machine learning: supervised, unsupervised, and reinforcement learning

· Difference between labeled and unlabeled data

· The difference between regression and classification, and how they are used

3. Linear Regression

· Fitting a line through a set of data points

· Coding the linear regression algorithm in Python

· Using Turi Create to build a linear regression model to predict housing prices in a real dataset

· What is polynomial regression

· Fitting a more complex curve to nonlinear data

· Examples of linear regression

4. Optimizing the Training Process

· What is underfitting and overfitting

· Solutions for avoiding overfitting

· Testing the model complexity graph, and regularization

· Calculating the complexity of the model

· Picking the best model in terms of performance and complexity

5. The perceptron Algorithm

· What is classification

· Sentiment analysis

· How to draw a line that separates points of two colors

· What is a perceptron

· Coding the perceptron algorithm in Python and Turi Create

6. Logistic Classifiers

· Hard assignments and Soft assignments

· The sigmoid function

· Discrete perceptrons vs. Continuous perceptrons

· Logistic regression algorithm for classifying data

· Coding the logistic regression algorithm in Python

7. Measuring Classification Models

· Types of errors a model can make

· The confusion matrix

· what are accuracy, recall, precision, F-score, sensitivity, and specificity

· what is the ROC curve

8. The Naive Bayes Model

· What is Bayes theorem

· Dependent and independent events

· The prior and posterior probabilities

· Calculating conditional probabilities

· using the naive Bayes model

· Coding the naive Bayes algorithm in Python

9. Decision Trees

· What is a decision tree

· Using decision trees for classification and regression

· Building an app-recommendation system using users’ information

· Accuracy, Gini index, and entropy

· Using Scikit-Learn to train a decision tree

10. Neural Networks

· What is a neural network

· Architecture of a neural network: nodes, layers, depth, and activation functions

· Training neural networks

· Potential problems in training neural networks

· Techniques to improve neural network training

· Using neural networks as regression models

Bonus Content / Time Permitting

11. Bonus: Support vector machine and the Kernel methods

· What a support vector machine

· Which of the linear classifiers for a dataset has the best boundary

· Using the kernel method to build nonlinear classifiers

· Coding support vector machines and the kernel method in Scikit-Learn

12. Bonus: Ensemble learning

· What ensemble learning is

· Using bagging to combine classifiers

· Using boosting to combine classifiers

· Ensemble methods: random forests, AdaBoost, gradient boosting, and XGBoost

13. Bonus: Real-World Example: Data Engineering and ML

· Cleaning up and preprocessing data to make it readable by our model

· Using Scikit-Learn to train and evaluate several models

· Using grid search to select good hyperparameters for our model

· Using k-fold cross-validation to be able to use our data for training and validation simultaneously