It remained for Nazi Germany to design the most satanic psychological experi ment of all time, the independent variables consisting of brutality, bestiality, physical and mental torture on an unprecedented scale. What were the effects of this massive assault on the human spirit, on man's ability to assimilate such experiences, if he survived physically? While the terror of the Nazi concentration camps has been indelibly engraved in the history of Western civilization as its most shameful chapter, little systematic study has been addressed to the subsequent lives of that minority of inmates who were fortunate enough to escape physical annihilation and lived to tell about their nightmare. Dr. PAUL MATUSSEK, a respected German psychiatrist, aided by a small group of collaborators, performed the task of identifying a group of victims (mostly Jews but also political prisoners), who, following their liberation, had settled in Germany, Israel, and the United States. By careful interviews, questionnaires, and psychological tests he brought to bear the methods of sensitive clinical inquiry on the experiences of those who dared to reminisce and who were sufficiently trusting to share their feelings and memories with clinical investigators. It is a telling commentary that many people, even after the passage of years, refused to respond."

Professor Bosch's study of infantile autism is a most valuable contribution to the slowly increasing body of knowledge about this baffling and most severe psychiatrie disorder of childhood. Reading it in the original German when it first appeared in 1962, I was greatly impressed by his deep sympathy for these unfortunate children and by his keen insight into the overt manifestations of a behavior which presents the observer with tantalizing riddles. Having spent nearly a lifetime in unravelling the meaning of the behavior of autistic children, I was much taken by Professor Bosch's very different approach to the same problem. His research sheds further light into the darkness that reigns in the mind of the autistic child. I am delighted that his important contribution is now easily available also to American readers. Everybody who works with children suffering from infantile autism for any length of time and also studies this disease, becomes impressed by how much their inability to relate and to resporrd appro"prrately can teach us about human psychology in general, and in particular how and why things go wrong in man's relations to his fellow man. All through his book, Professor Bosch correctly stresses that autistic behavior is neither asymptom nor a syndrome, but a unique form of breakdown in all inter personal relations."

The formalism of probabilistic graphical models provides a unifying framework for capturing complex dependencies among random variables, and building large-scale multivariate statistical models. Graphical models have become a focus of research in many statistical, computational and mathematical fields, including bioinformatics, communication theory, statistical physics, combinatorial optimization, signal and image processing, information retrieval and statistical machine learning. Many problems that arise in specific instances-including the key problems of computing marginals and modes of probability distributions-are best studied in the general setting. Working with exponential family representations, and exploiting the conjugate duality between the cumulant function and the entropy for exponential families, Graphical Models, Exponential Families and Variational Inference develops general variational representations of the problems of computing likelihoods, marginal probabilities and most probable configurations. It describes how a wide variety of algorithms- among them sum-product, cluster variational methods, expectation-propagation, mean field methods, and max-product-can all be understood in terms of exact or approximate forms of these variational representations. The variational approach provides a complementary alternative to Markov chain Monte Carlo as a general source of approximation methods for inference in large-scale statistical models.