series5 - Exercises Machine Learning Laboratory Dept of...

Info icon This preview shows pages 1–2. Sign up to view the full content.

Machine Learning Laboratory Dept. of Computer Science, ETH Z¨urich Prof. Dr. Joachim M. Buhmann Web Email questions to: Dmitry Laptev [email protected] Exercises Machine Learning AS 2012 Series 5, Nov 21th, 2012 (Adaboost) Problem 1 (Combining Kernels): Recall the definition of positive semidefiniteness. eigenvalues Show that positive semidefinite matrices have non-negative eigenvalues. In this exercise, we explore several common ways to combine known kernels so that we obtain new kernels. For some input data x, x 0 ∈ X , let k 1 ( x, x 0 ) and k 2 ( x, x 0 ) be kernels over X × X → R . Most of the following proofs can be done by giving the corresponding mapping Φ to feature space. scaling For a positive real number a , show that k ( x, x 0 ) := ak 1 ( x, x 0 ) is a kernel. sum Show that k ( x, x 0 ) := k 1 ( x, x 0 ) + k 2 ( x, x 0 ) is a kernel. linear combination For real positive numbers a and b , show using the results before that k ( x, x 0 ) := ak 1 ( x, x 0 )+ bk 2 ( x, x 0 ) is a kernel. product Show that k ( x, x 0 ) := k 1 ( x, x 0 ) k 2 ( x, x 0 ) is a kernel. exponentiated Show that for a positive integer p , k ( x, x 0 ) := k 1 ( x, x 0 ) p is a kernel metric For x, x 0 R d , and for a positive definite matrix M , show that k ( x, x 0 ) := x > Mx 0 is a kernel. Problem 2 (Deriving the SVM Dual): Recall the soft margin C -SVM: minimize w,b,ξ 1 2 k w k 2 + C n X i =1 ξ i (1) subject to y i ( h w, x i i + b ) 1 - ξ i for all i = 1 , . . . , n. (2) ξ i 0 for all i = 1 , . . . , n (3) 1. Write down the Lagrangian, using α i as the Lagrange multiplier corresponding to constraint (2) and γ i as the Lagrange multiplier corresponding to constraint (3).
Image of page 1

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

Image of page 2
This is the end of the preview. Sign up to access the rest of the document.
  • Fall '12
  • J. M. Buhmann
  • Machine Learning, WI, Boosting, AdaBoost, Positive-definite matrix, AdaBoost algorithm, base classifier

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern