25F - DL Theory

EECE695D: Deep Learning Theory (Fall 2025) #

Team #

  • Instructor. Jaeho Lee 이재호 ✉️
  • TA. Byung-ki Kwon 권병기 ✉️

Location & Time #

  • Class. Mondays/Wednesdays, 2PM–3:30PM, EB2, #109
  • Office Hr. Mondays, 5PM–6PM, GoAround Coffee @RIST (or by request)

Schedule (tentative) #

  • W1. Overview and Basics of Statistical Learning Theory (9/1, 9/3)
  • W2. Approx: Universal approximation with shallow nets (9/8, 9/10)
  • W3. Approx: Infinite-width and kernels (9/15)
  • W4. Approx: Benefits of depth
  • W7. Optim: Convex optimization and generalizations
  • W6. Chuseok Holidays
  • W8. Optim: SGD and flow-based analyses
  • W9. Mid-Term
  • W10. Optim: Implicit bias
  • W11. Gen: Concentration of measures, uniform convergence
  • W12. Gen: Rademacher complexities, covering numbers
  • W13. Gen: Chaining and VC dimensions
  • W14. Recent Results
  • W15. Student Presentations - 1
  • W16. Student Presentations - 2

Textbook #