Why machine learning and AI?; theoretical foundations for machine learning; representation of complex knowledge by clauses; representation of knowledge about actions; learning by doing - LEX, SAGE; the version spaces - a formal presentation; explanation-based learning - Mitchell's and DeJong's views of EBL; empirical learning by detection of similarities - a detailed example of Michalski's approach; rational learning by similarity detection - the "rational" approach; automatic construction of taxonomies; debugging and deep understanding; learning by analogy - Winston's approach to learning by analogy. Appendices: theoretical grounds of the equivalence between theorems and clauses - Herbrand's universe of a set of clauses, skolemization, semantic trees; synthesis of predicates - the role of program synthesis in ML, program synthesis from input-output sequences; ML in context - epistemological and social issues.