Why aren't the most powerful new technologies being used to solve the world's most important problems: hunger, poverty, conflict, employment, disease? In Link, Dr. Lorien Pratt answers these ...questions by exploring the solution that is emerging worldwide to take Artificial Intelligence to the next level: Decision Intelligence.
The U.S. government spends enormous resources each year on the gathering and analysis of intelligence, yet the history of American foreign policy is littered with missteps and misunderstandings that ...have resulted from intelligence failures. In Why Intelligence Fails, Robert Jervis examines the politics and psychology of two of the more spectacular intelligence failures in recent memory: the mistaken belief that the regime of the Shah in Iran was secure and stable in 1978, and the claim that Iraq had active WMD programs in 2002. The Iran case is based on a recently declassified report Jervis was commissioned to undertake by CIA thirty years ago and includes memoranda written by CIA officials in response to Jervis's findings. The Iraq case, also grounded in a review of the intelligence community's performance, is based on close readings of both classified and declassified documents, though Jervis's conclusions are entirely supported by evidence that has been declassified. In both cases, Jervis finds not only that intelligence was badly flawed but also that later explanations—analysts were bowing to political pressure and telling the White House what it wanted to hear or were willfully blind—were also incorrect. Proponents of these explanations claimed that initial errors were compounded by groupthink, lack of coordination within the government, and failure to share information. Policy prescriptions, including the recent establishment of a Director of National Intelligence, were supposed to remedy the situation. In Jervis's estimation, neither the explanations nor the prescriptions are adequate. The inferences that intelligence drew were actually quite plausible given the information available. Errors arose, he concludes, from insufficient attention to the ways in which information should be gathered and interpreted, a lack of self-awareness about the factors that led to the judgments, and an organizational culture that failed to probe for weaknesses and explore alternatives. Evaluating the inherent tensions between the methods and aims of intelligence personnel and policymakers from a unique insider's perspective, Jervis forcefully criticizes recent proposals for improving the performance of the intelligence community and discusses ways in which future analysis can be improved.
Boosting Schapire, Robert E; Freund, Yoav
2012, 2012-05-18, c2012
eBook
Odprti dostop
Boosting is an approach to machine learning based on the idea of creating a highly accurate predictor by combining many weak and inaccurate "rules of thumb." A remarkably rich theory has evolved ...around boosting, with connections to a range of topics, including statistics, game theory, convex optimization, and information geometry. Boosting algorithms have also enjoyed practical success in such fields as biology, vision, and speech processing. At various times in its history, boosting has been perceived as mysterious, controversial, even paradoxical. This book, written by the inventors of the method, brings together, organizes, simplifies, and substantially extends two decades of research on boosting, presenting both theory and applications in a way that is accessible to readers from diverse backgrounds while also providing an authoritative reference for advanced researchers. With its introductory treatment of all material and its inclusion of exercises in every chapter, the book is appropriate for course use as well. The book begins with a general introduction to machine learning algorithms and their analysis; then explores the core theory of boosting, especially its ability to generalize; examines some of the myriad other theoretical viewpoints that help to explain and understand boosting; provides practical extensions of boosting for more complex learning problems; and finally presents a number of advanced theoretical topics. Numerous applications and practical illustrations are offered throughout.
