Algorithms that learn from data Traditional programming takes data and rules and produces an output. Machine learning takes data and the desired output and produces rules, the model weights w. In the final view, training data and desired output go into ML training, which produces the rules; the rules and test data go into ML testing, which produces the output. algorithms that can learn from data Training: learns the rules from examples Testing: applies the learned rules, like traditional programming rules move from input to output data test data rules traditionalprogramming ML testing output data training data desired output ML ML training rules ML model weights w