Practical Machine Learning

Instructor: Alejandro Correa Bahnsen

• email: al.bahnsen@gmail.com
• twitter: @albahnsen
• github: albahnsen

The use of statistical models in computer algorithms allows computers to make decisions and predictions, and to perform tasks that traditionally require human cognitive abilities. Machine learning is the interdisciplinary field at the intersection of statistics and computer science which develops such algorithnms and interweaves them with computer systems. It underpins many modern technologies, such as speech recognition, internet search, bioinformatics, computer vision, Amazon’s recommender system, Google’s driverless car and the most recent imaging systems for cancer diagnosis are all based on Machine Learning technology.

This course on Machine Learning will explain how to build systems that learn and adapt using real-world applications. Some of the topics to be covered include machine learning, python data analysis, deep learning frameworks, natural language processing models and recurrent models. The course will be project-oriented, with emphasis placed on writing software implementations of learning algorithms applied to real-world problems, in particular, image analysis, image captioning, natural language pocessing, sentiment detection, among others.

Requiriments

• Python version 3.5;
• Numpy, the core numerical extensions for linear algebra and multidimensional arrays;
• Scipy, additional libraries for scientific programming;
• Matplotlib, excellent plotting and graphing libraries;
• IPython, with the additional libraries required for the notebook interface.
• Pandas, Python version of R dataframe
• Seaborn, used mainly for plot styling
• scikit-learn, Machine learning library!

A good, easy to install option that supports Mac, Windows, and Linux, and that has all of these packages (and much more) is the Anaconda.

GIT!! Unfortunatelly out of the scope of this class, but please take a look at these tutorials

Evaluation

• 30% Exercises
• 50% Projects
• 20% Final Project

Schedule

Supervised Machine Learning

Date	Session	Notebooks/Presentations	Exercises
January 21st	Introduction to python and ML	1 - Intro to ML 2 - Intro to Python for data analysis	E1 - Data Science Overview E2 - Python Text Analysis
January 28th	Linear Models	3 - Linear Regression 4 - Logistic Regression	E3 - Supervised vs Unsupervised Overview E4 - Linear Models
February 4th	SVM & Decision Trees	5 - SVM 6 - Decision Tress 7 - Regularization	E5 - Decision Trees Overview E6 - SVM - Regularization E7 - Decision Trees
February 11th	Machine Learning as a Service	8 - Introduction to APIs 9 - Model Deployment 10 - Creating APIs in AWS	P1 - Survival Prediction API
February 18th	Ensembles	11 - Bagging 12 - Boosting	E8 - Best Ensemble Overview E9 - Bagging
February 25th	Random Forest	13 - Random Forest	E10 - Random Forest Performance Review E11 - Random Forest
March 4th	Feature Engineering	14 - Feature Engineering	E12 - Gradient Boosting Review E13 - Categorical Encoding
March 11th	Project Presentations
March 18th	Unbalanced Learning	15 - Unbalanced Learning	E14 - Unbalanced Learning Overview E15 - Fraud Detection

Natural Language Processing

Date	Session	Notebooks/Presentations	Exercises
April 1st	Natural Language Processing	16 - Introduction to NLP 17 - Introduction to NLP II	E16 - NLP Review E17 - Sentiment Analysis P2 - NLP Movies Analysis API
April 8th	Sentiment Analysis	18 - Sentiment Analysis	E18 - Homeworks Analysis
April 22nd	Project Presentations

Advanced Topics in Machine Learning

Date	Session	Notebooks/Presentations	Exercises
April 29th	Introduction to Deep Learning	19 - Introduction to Neural Networks - MLP 20 - Deep Learning in keras	E19 - Neural Networks Fundamentation E20 - Neural Networks in Keras
May 6th	Introduction to Deep Learning II	20 - LSTM 21 - CNN

Final Project Presentation

• P3 - Kaggle Competition