2018-1 Optimization
Time: Wed 9:00am-12:00
Location: Cluster Bd. R509 (학연산클러스터 509호)
Textbook: Numerical Optimization, Nocedal & Wright, Springer, 2006
Grading:
- Homework: 30%
- Midterm Exam: 30%
- Final Exam 30%
- Attendance: 10%
June 15: homework submission: due on June 15 by 13:00 @ Cluster Bd R. 620 (학연산클러스터 620호 인공지능연구실)
June 20: Final Exam
Latex template for scribblers: [tex]
Lecture Notes
- Review of analysis (pdf, scribbler: 이정현)
- Rate of convergence, descent direction (pdf, scribbler: 김영석)
- Differentiation, continuity(pdf, scribbler: 손수현)
- HW1 [pdf] : due April 4 (no submission, volunteer-based discussion in class. Updated March 28)
- Optimality conditions (pdf, scribbler: 권준형)
- Convex optimization (pdf, scribbler: 이정현)
- HW2 [pdf]: due April 18 (no submission, volunteer-based discussion in class.)
- Convex Optimization (pdf, scribbler: 강동연)
- Lipschitz continuity, strong convexity
- Midterm exam: April 25 (in class)
- You can bring one A4 paper, filling up only ONE-SIDE of the paper with the content from the lecture. This is optional
- You can bring your own scratch paper
- Exam questions will be at the similar level to the homework questions, but may not be the same
- Steepest descent & Newton's method (pdf, scribbler: M. Ibtesam)
- Convergence Rate of GD & Newton's method, Quasi-Newton Method (pdf, scribbler: 손수현)
- Quasi-Newton Method, Conjugate Gradient Method (pdf, scribbler: Z. Wu)
- HW3 [pdf]: due May 30
- May 23: Conjugate gradient (pdf, scribbler: 이기찬), SGD (link)
- May 30: PGD (link), AGD (link)
- June 5 (make-up lecture): 1:30 ~ 4:30 @ Cluster R.506
- June 6: no lecture (memorial day)
- June 13: no lecture (election day)
- June 15: homework submission: due on June 15 by 13:00 @ Cluster Bd R. 620 (학연산클러스터 620호 인공지능연구실)
- June 20: Final Exam