No-Free-Lunch Theorem
It is shown that no learning algorithm can learn every function in the PAC model, and a characterization of what hypothesis classes are learnable in the PAC model is built up.