Two SVM tutorials linked in class website (please, read both): High-level presentation with applications (Hearst 1998) Detailed tutorial (Burges 1998)SVMs, Duality and the Kernel TrickMachine Learning – 10701/15781Carlos GuestrinCarnegie Mellon UniversitythFebruary 27 , 20062006 Carlos Guestrin 1Announcements Third homework is outst Due March 1 Final assigned by registrar: May 12, 1-4p.m Location TBD Midtermth March 8 , a week from Wednesday Open book, notes, papers, etc. No computers22006 Carlos GuestrinSVMs reminder32006 Carlos GuestrinToday’s lecture Learn one of the most interesting and exciting recent advancements in machine learning The “kernel trick” High dimensional feature spaces at no extra cost! But first, a detour Constrained optimization!42006 Carlos GuestrinConstrained optimization52006 Carlos GuestrinLagrange multipliers – Dual variablesMoving the constraint to objective functionLagrangian:Solve:62006 Carlos GuestrinLagrange multipliers – Dual variablesSolving:72006 Carlos GuestrinDual SVM derivation (1) –the linearly separable case82006 Carlos GuestrinDual SVM derivation (2) –the linearly separable case92006 Carlos GuestrinDual SVM interpretation102006 Carlos Guestrinw.x+ b = 0Dual SVM formulation –the linearly separable case112006 Carlos GuestrinDual SVM derivation –the non-separable case122006 Carlos GuestrinDual SVM formulation –the ...
Two SVM tutorials linked in class website (please, read both): High-level presentation with applications (Hearst 1998) Detailed tutorial (Burges 1998)
SVMs, Duality and the Kernel Trick
Machine Learning – 10701/15781 Carlos Guestrin Carnegie Mellon University
February 27th, 2006 2006 Carlos Guestrin
Announcements
Third homework is out Due March 1st
Final assigned by registrar: May 12, 1-4p.m Location TBD
Midterm March 8th, a week from Wednesday Open book, notes, papers, etc. No computers
2006 Carlos Guestrin
SVMs
reminder
2006 Carlos Guestrin
Today’s lecture
Learn one of the most interesting and exciting recent advancements in machine learning The “kernel trick” High dimensional feature spaces at no extra cost! But first, a detour Constrained optimization!
2006 Carlos Guestrin
Constrained optimization
2006 Carlos Guestrin
Lagrange multipliers – Dual variables
Solve:
Moving the constraint to objective function Lagrangian:
2006 Carlos Guestrin
Lagrange multipliers –
Solving:
2006 Carlos Guestrin
Dual variables
Dual SVM derivation (1) – the linearly separable case
2006 Carlos Guestrin
Dual SVM derivation (2) – the linearly separable case
2006 Carlos Guestrin
Dual SVM interpretation
2006 Carlos Guestrin
Dual SVM formulation – the linearly separable case