SP11 cs188 lecture 22 -- kNN kernels ++ 6PP

SP11 cs188 lecture 22 -- kNN kernels ++ 6PP - Announcements...

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

View Full Document Right Arrow Icon
CS 188: Artificial Intelligence Spring 2010 Lecture 22: Nearest Neighbors, Kernels 4/18/2011 Pieter Abbeel – UC Berkeley Slides adapted from Dan Klein Announcements s On-going: contest (optional and FUN!) s Remaining lectures: s Today: Machine Learning: Nearest Neighbors, Kernels s Wednesday: Machine Learning for Computer Vision s Next Monday: Case Studies in Speech/Language and Robotics s Next Wednesday: s Course Wrap-Up s Pointers to courses and Books for those who want to learn more AI s Contest! s RRR Week Monday and Wednesday: Review Sessions Today s Nearest neighbors s Kernels s Applications: s Extension to ranking / web-search s Pacman apprenticeship Classification Hello, Do you want free printr cartriges? Why pay more when you can get them ABSOLUTELY FREE! Just # free : 2 YOUR_NAME : 0 MISSPELLED : 2 FROM_FRIEND : 0 ... SPAM or + PIXEL-7,12 : 1 PIXEL-7,13 : 0 ... NUM_LOOPS : 1 ... “2” Classification overview s Naïve Bayes: s Builds a model training data s Gives prediction probabilities s Strong assumptions about feature independence s One pass through data (counting) s Perceptron: s Makes less assumptions about data s Mistake-driven learning s Multiple passes through data (prediction) s Often more accurate s MIRA: s Like perceptron, but adaptive scaling of size of update s SVM: s Properties similar to perceptron s Convex optimization formulation s Nearest-Neighbor: s Non-parametric: more expressive with more training data s Kernels s Efficient way to make linear learning architectures into nonlinear ones Case-Based Reasoning s Similarity for classification s Case-based reasoning s Predict an instance’s label using similar instances s Nearest-neighbor classification s 1-NN: copy the label of the most similar data point s K-NN: let the k nearest neighbors vote (have to devise a weighting scheme) s Key issue: how to define similarity s Trade-off: s Small k gives relevant neighbors s Large k gives smoother functions s Sound familiar? s
Background image of page 1

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

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 08/26/2011 for the course CS 188 taught by Professor Staff during the Spring '08 term at University of California, Berkeley.

Page1 / 6

SP11 cs188 lecture 22 -- kNN kernels ++ 6PP - Announcements...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online