ECE 595ML Lecture 5.1: Gradient Descent

By Stanley H. Chan

Electrical and Computer Engineering, Purdue University, West Lafayette, IN

View Presentation

Category

Online Presentations

Published on

Cite this work

Researchers should cite this work as follows:

  • Stanley H. Chan (2020), "ECE 595ML Lecture 5.1: Gradient Descent," https://nanohub.org/resources/32215.

    BibTex | EndNote

Time

Location

WTHR 200, Purdue University, West Lafayette, IN

ECE 595ML Lecture 5.1: Gradient Descent
  • Lecture 5.1: Gradient Descent 1. Lecture 5.1: Gradient Descent 0
    00:00/00:00
  • Outline 2. Outline 44.711378044711381
    00:00/00:00
  • Outline 3. Outline 141.27460794127461
    00:00/00:00
  • Gradient Descent 4. Gradient Descent 185.68568568568568
    00:00/00:00
  • Why is the direction − f (x)? 5. Why is the direction − f (x)… 437.70437103770439
    00:00/00:00
  • Descent Direction 6. Descent Direction 535.1351351351351
    00:00/00:00
  • The Steepest d 7. The Steepest d 674.60794127460792
    00:00/00:00
  • Steepest Descent Direction 8. Steepest Descent Direction 902.10210210210209
    00:00/00:00
  • Step Size 9. Step Size 1050.6172839506173
    00:00/00:00
  • Convergence 10. Convergence 1404.9382716049383
    00:00/00:00
  • Understanding Convergence 11. Understanding Convergence 1557.8578578578579
    00:00/00:00
  • Advice on Gradient Descent 12. Advice on Gradient Descent 1716.8501835168502
    00:00/00:00
Pop Out