This lecture summarizes the general result established for linear regression, focusing on the bound relating the risk of predictors to the empirical risk, considering the rate of macro complexity and high probability statements. It also introduces the concept of Vapnik-Chervonenkis dimension and growth function.