Here is a plan to learn ML fundamentals in an afternoon by watching some videos on youtube:

Follow this plan

Machine learning fundamentals:

• What is Bias and Variance? – 6 minutes
• What is a Tensor? – 12 minutes

[Stop and drink coffee, eat a snack]

How to address bias and variance:

• Regularization (e.g. L2/Ridge) – 20 minutes
• Bagging (e.g. Random forrest) – 9 minutes
• Boosting (e.g. AdaBoost) – 20 minutes
• L1 vs L2 regularization – 9 minutes

Extra material:

• Bootstrapping – 9 minutes

Test your knowledge

• What is bias?
  • A: Bla
  • B: The inability of a machine learning model (e.g. linear regression) to express the true relationship between X and Y
  • C: Bla
• What is variance?
  • A: The difference in how well a model fits different datasets (e.g. training and test)
  • B: Bla
  • C: Bla
• What problem does regularization, bagging and boosting address?
  • A: Bla
  • B: Bla
  • C: Finds the sweet spot between simple and complicated models
• What is regularization?
  • A: Bla
  • B: Bla
  • C: Bla
• What is bagging?
  • A: Bla
  • B: Bla
  • C: Bla
• What is boosting?
  • A: Bla
  • B: Bla
  • C: Bla
  • What is bootstrapping?
  • A: Repeat an experiment a bunch of times until we feel certain about the result
  • B: Repeatly random sample n times (with replacement) from a set of n observations and build up a histogram of any statistic, e.g. the mean.
  • C: Augment a small set of observations with synthetic samples to increase sample size