![[Home]](/ailab/people/andel/pics/AIWiki.png)
The SupervisedLearningTask is to use examples to train a system how to behave. More specifically, you want the system to output the right thing when it is given an input of a certain form. You have a bunch of examples of inputs and the correct outputs to go with them. You want the system to look at the examples, and learn from them how to compute the correct outputs for this sort of input.
The examples are called “TrainingData”. You know what the right answers are for the TrainingData; this is called a “TargetFunction?”. The TargetFunction? which evaluates the rightness of any output when the TrainingData is given as input.
Steps in solving a SupervisedLearningTask:
1. You choose a TargetFunction? and find some TrainingData.
2. You pick a [[Representation?]].
3. You pick a learning algorithm for that [[Representation?]].
4. You run the learning algorithm on the TrainingData, producing a model.
5. You run the model on TestData? to evaluate its performance.