algorithms · ML · 2017-03-06 · yuex

I came across a paper called A Few Useful Things to Know about Machine Learning which summarized some black arts in machine learning.

  1. Learning = Representation + Evaluation + Optimization
  2. It's generalization that counts
  3. Data alone is not enough (we also need knowledge)
  4. Overfitting has many faces
  5. Intuition fails in high dimensions
  6. Theoretical guarantee are not what they seem
  7. Feature engineering is the key
  8. More data beats a clever algorithm
  9. Learn many models, not just one (model ensembles)
  10. Simplicity does not imply accuracy
  11. Representable does not imply learnable
  12. Correlation does not imply causation

The author also recommended his book The Master Algorithm and his ML course.