The machine learning (ML) space can seem intimidating — where do you start, and how does it all connect? At Manifold, we've compiled a list of resources borne of our expertise that we think will help you form a strong ML foundation. We hope to continue updating and improving this list, and may reshare it out periodically in the hopes that those who embark on this journey can have an even smoother and more fulfilling experience.
Some of these topics may already be familiar to you and some less so; feel free to skip through things you already know. And let us know in the Comments below what else you would add to the list. Enjoy!
There are three sections to this content
- Business-oriented content for executives about ML (generally quick reads)
- Technical content about ML (think survey papers)
- Hands-on-keyboard courses/projects (tinkering for an hour to a few days per project)
1. Business-oriented content for executives about ML
- McKinsey piece on Exec's guide to AI
- This piece does a nice job of explaining why we are seeing an explosion in machine learning now, the major categories of machine learning including deep learning, and the classes of business problems that are amenable to each approach.
- Dataiku whitepaper on different types of ML
- This piece, while a bit heavy on the marketing (it's by a product company) does explain in more technical detail some of the machine learning approaches and the pros/cons of each.
- Ways to think about machine learning — Ben Evans
- Framework for thinking about ML problems: Known data, known problems; known data, new questions; new data, new questions.
- AI for the Real World — Tom Davenport & Rajeev Ronanki
- While the article is heavy on RPA (a Deloitte area of focus), it makes the nice distinction between moonshot projects and incremental wins like IT staffing, with MD Anderson as a case study.
- AI Transformation Playbook
- Andrew Ng (of quite a bit of fame) is now part of an AI consulting company also. He wrote up a nice piece about AI Transformation at large organizations.
Somewhat more technical, but not totally technical
- Rist Siilasmaa Chairman of Nokia on Machine Learning (video, 1 hour)
- Why Machine Learning is Hard (N^3 search space) — S. Zayd Enam
- How to start with AI — Sourav Dey
- The importance of baseline models — Emmanuel Ameisen
- AI Hierarchy of Needs — Monica Rogati
- Why businesses fail at machine learning — Cassie Kozyrkov
- The difference between machine learning research and product development with ML
2. Technical Content about ML
- A Gentle Introduction to Transfer Learning for Deep Learning — Jason Brownlee
- The Random Forest Algorithm — Niklas Donges
- Gradient Boosting from Scratch — Prince Grover
- Deep Learning for Videos: A 2018 Guide to Action Recognition — Rohit Ghosh
- Great survey paper on the state of video recognition and why it's hard.
- Machine Learning: The High-Interest Credit Card of Technical Debt — Google
3. Hands-on-Keyboard courses/projects
- ML crash course — Google (about one week)
- Fast AI
- Two courses: "practical deep learning for coders" and "cutting edge deep learning for coders"
- Neural Networks and Deep Learning
- In-depth guide for how to build a neural network for recognizing handwritten digits using the MNIST data set. Free online book with Python code.
- AWS ML blog
- Has several posts that are actually examples, including data and code to walk you through. For example: Create a model for predicting orthopedic pathology using Amazon SageMaker
- Video Series on Neural Networks — 3Blue1Brown
- Essence of Linear Algebra and Essence of Calculus — 3Blue1Brown
- High-quality course, and super helpful if you forgot all the Math you learned in college, or never had a chance to learn it in the first place.