CS229 - Lecture 1
Taking Andrew Ng’s machine learning class at Stanford this quarter. Here are some memorable points from the first lecture.
Definition of machine learning: Program learns from experience E on some performance measure P if P improves with experience E.
“SUPERVISED LEARNING”
Given set of example and correct answers, build a predictor that estimates new datasets
Examples:
- regression (house prices)
- classification/labeling (spam)
“UNSUPERVISED LEARNING”
given a dataset - find structure (k means/clustering)
examples image segmentation, cocktail party problem (ICA)
“REINFORCEMENT LEARNING”
use reward/punishment methodology to perform some task - example robots learning to walk