What is knowledge and how is it represented? This book focuses on the idea of formalising knowledge as relations, interpreting knowledge represented in databases or logic programs as relational data and discovering new knowledge by identifying hidden and defining new relations. After a brief introduction to representational issues, the author develops a relational language for abstract machine learning problems. He then uses this language to discuss traditional methods such as clustering and decision tree induction, before moving onto two previously underestimated topics that are just coming to the fore: rough set data analysis and inductive logic programming. Its clear and precise presentation is ideal for undergraduate computer science students. The book will also interest those who study artificial intelligence or machine learning at the graduate level. Exercises are provided and each concept is introduced using the same example domain, making it easier to compare the individual pro
This is a collection of lecture notes from the Summer School 'Cycles Algébriques; Aspects Transcendents, Grenoble 2001'. The topics range from introductory lectures on algebraic cycles to more advanced material. The advanced lectures are grouped under three headings: Lawson (co)homology, motives and motivic cohomology and Hodge theoretic invariants of cycles. Among the topics treated are: cycle spaces, Chow topology, morphic cohomology, Grothendieck motives, Chow-Künneth decompositions of the diagonal, motivic cohomology via higher Chow groups, the Hodge conjecture for certain fourfolds, an effective version of Nori's connectivity theorem, Beilinson's Hodge and Tate conjecture for open complete intersections. As the lectures were intended for non-specialists many examples have been included to illustrate the theory. As such this book will be ideal for graduate students or researchers seeking a modern introduction to the state-of-the-art theory in this subject.