Home Curriculum Learning Log Bookshelf

Curriculum – Machine Learning Sabbatical

Constructed in collaboration with a-martyn

Summary

• Time: 3 months full-time
• Prequesites: Python programming, A-level / intermediate undergraduate level maths

This curriculum aims to provide a 3-month course teaching the prerequisite mathematics and foundational principles of machine learning. We have opted for a bottom-up approach focussing on fundamental principles whilst also allowing scope for applied experimentation.

Three tracks of study are pursued in parallel:

1. Foundational track: Structured course-based study of fundamental concepts that apply broadly across machine learning including mathematics, statistics and core learning algorithms.
2. Applied track: Goal driven projects intended to test understanding and application of knowledge, and build strong communication of machine learning concepts
3. Practice track: A "little and often" track; small programming exercises and mathematical problems. Intended to keep maintain and enhance practical skills.

1. Foundational Track

Introduction to Machine Learning (60 hrs)

• Machine Learning by Andrew Ng

Mathematics

• Course: Mathematics for Machine Learning (60 hrs)
• Textbook: DS-GA 1002: Statistical and Mathematical Methods
• Checklist
• Cribsheet

Supervised Learning & Statistics (90 hrs)

• Course: Foundations of Machine Learning by Bloomberg ML EDU
• Textbooks:
  • Hands-On Machine Learning with Scikit-Learn and TensorFlow (Aurélien Géron)
  • The Elements of Statistical Learning (Hastie, Friedman, and Tibshirani)
  • An Introduction to Statistical Learning
  • Understanding Machine Learning: From Theory to Algorithms (Shalev-Shwartz and Ben-David)
  • Machine Learning: a Probabilistic Perspective by Kevin Murhy
  • Data Science for Business (Provost and Fawcett)

UPDATE

Supervised Learning & Statistics (60 hrs)

• Course: Statistical Learning by Trevor Hastie and Rob Tibshirani
• Textbooks:
  • The Elements of Statistical Learning (Hastie, Friedman, and Tibshirani)
  • An Introduction to Statistical Learning

Unsupervised Learning / Graphical Models

• Course: TBD
• Textbook: TBD

2. Applied track

• Build a dataset from scratch
• Look for problems where dataset can be built quickly
• 20% increasing to 50% of course time as knowledge progresses
• Projects should be fun, and have toy-like quality
• Focus on demonstrable deliverables (1 per week)

3. Practice track

Software Engineering

• Course: InterviewCake.com
• Textbook: Cracking the coding interview

Mathematics

Linear Algebra

• Questions: Practice Questions
• Course: Pavel Grinfeld's Lemma

Caclulus

• TBD

Probability and Statistics

• TBD

Time commitments

• 60% Foundational Track
• 20% Applied Track
• 20% Practice Track
• 1 Meetup / Week

Desired outcomes

• [x] Machine Learning by Andrew Ng coursera accreditation.
• [x] Mathematics for Machine Learning: Linear Algebra accreditation.
• [x] Mathematics for Machine Learning: Multivariate Calculus accreditation.
• [x] Mathematics for Machine Learning: PCA accreditation.
• [ ] Completed Homeworks from Foundations of Machine Learning published as Jupyter notebooks.
• [ ] 12 Jupyter notebooks demonstrating fundamental topics in machine learning.
• [ ] At least one demonstrable applied ML project.
• [ ] An offer of employment :)