Monty Hall a b c *(Goat not necessarily behind Door b) 3 doors: a, b, c behind two: Goat behind one: Prize You pick one door, but are not shown the contents Host opens one of the other two doors that has a Goat You now have the option to switch to the other unopened door Should you switch? *(Goat not necessarily behind Door b) a b c
Monty Hall in Vermont
Bayesian Inference in AI Medical Diagnosis: Pr(positive | age, sex, health history, symptoms) Game Playing: Pr(get black jack | cards dealt) Credit Analysis: Pr(good credit | credit history) Stock Purchasing: Pr(stock will go up | current price, history, trends)
? Recall: Supervised Learning b a a b a Given new example, put it into a group.
Training Classifying Learned model/ Classifier Training Set Extract features/ labels Train Decision Trees Bayesian Learning Neural Nets... Classifying Learned model/ Classifier Label Instance/Example Extract features
Tennis example Training Set: New Example: Overcast, Cool, Normal, Strong Play Tennis?
Training Classifying Learned model/ Classifier Training Set Extract features/ labels Train Decision Trees Bayesian Learning Neural Nets... Classifying Learned model/ Classifier Label Instance/Example Extract features
Extracting Features/Labels: Training Raw data Label features Label f1, f2, f3, …, fn f1, f2, f3, …, fn Train f1, f2, f3, …, fn 1 1 f1, f2, f3, …, fn 1 extract features 1 f1, f2, f3, …, fn Training examples
Extracting Features: Testing/Classiftying Raw data labels features 1 f1, f2, f3, …, fn Learned model/ Classifier f1, f2, f3, …, fn f1, f2, f3, …, fn 1 extract features f1, f2, f3, …, fn predict the label f1, f2, f3, …, fn
Bayesian Learning Naïve Bayes Classifier