The Modern Mathematics of Deep Learningexplained for non-mathematicians

The curse of dimensionality

High-dimensional problems can become expensive. The source studies additional assumptions that can help.

Lessons

  1. Grid size
  2. Sample average
  3. Sampling a PDE representation
  4. Barron approximation bound
  5. Combining local descriptions
  6. A one-dimensional PDE example
  7. A sampled threshold feature
  8. A deterministic process in a PDE representation
  9. A function defined by expectation
  10. A sampling error bound
  11. Fourier representation and sampling weights
  12. A problem in the printed representation
  13. Approximating a random quantity by steps
  14. A function on a subset
  15. Expectation of a threshold feature
  16. Smooth local functions
  17. Support of a local weight
  18. A Fourier frequency bound
  19. Random input for a PDE example