"Learning and Generalisation" is a scholarly book that presents a formal mathematical theory addressing fundamental questions of machine learning. It explores how machines learn concepts from examples, how neural networks can make predictions for previously unknown inputs after training, and how much training is required to achieve a certain level of accuracy. Additionally, it discusses the identification of the dynamic behavior of nonlinear control systems through the observation of their input-output behavior over a limited time period. The second edition of the book expands the topics to include new areas such as support vector machines and the application of fat-shattering dimensions to the learning of neural networks. It also provides solutions to some of the open problems raised in the first edition and introduces new challenges.