From the Publisher:
Probabilistic Reasoning in Intelligent Systems is a complete andaccessible account of the theoretical foundations and computational methods that underlie plausible reasoning under uncertainty. The author provides a coherent explication of probability as a language for reasoning with partial belief and offers a unifying perspective on other AI approaches to uncertainty, such as the Dempster-Shafer formalism, truth maintenance systems, and nonmonotonic logic. The author distinguishes syntactic and semantic approaches to uncertaintyand offers techniques, based on belief networks, that provide a mechanism for making semantics-based systems operational. Specifically, network-propagation techniques serve as a mechanism for combining the theoretical coherence of probability theory with modern demands of reasoning-systems technology: modular declarative inputs, conceptually meaningful inferences, and parallel distributed computation. Application areas include diagnosis, forecasting, image interpretation, multi-sensor fusion, decision support systems, plan recognition, planning, speech recognitionin short, almost every task requiring that conclusions be drawn from uncertain clues and incomplete information.
Probabilistic Reasoning in Intelligent Systems will be of special interest to scholars and researchers in AI, decision theory, statistics, logic, philosophy, cognitive psychology, and the management sciences. Professionals in the areas of knowledge-based systems, operations research, engineering, and statistics will find theoretical and computational tools of immediate practical use. The book can also be used as an excellent text for graduate-level courses in AI, operations research, or applied probability.

Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference

Many areas of science depend on exploratory data analysis and visualization. The need to analyze large amounts of multivariate data raises the fundamental problem of dimensionality reduction: how to discover compact representations of high-dimensional data. Here, we introduce locally linear embedding (LLE), an unsupervised learning algorithm that computes low-dimensional, neighborhood-preserving embeddings of high-dimensional inputs. Unlike clustering methods for local dimensionality reduction, LLE maps its inputs into a single global coordinate system of lower dimensionality, and its optimizations do not involve local minima. By exploiting the local symmetries of linear reconstructions, LLE is able to learn the global structure of nonlinear manifolds, such as those generated by images of faces or documents of text.

/pdf/nonlinear-dimensionality-reduction-by-locally-linear-vmiazg4oee.pdf

Nonlinear dimensionality reduction by locally linear embedding.

From the Publisher:
With this text, you gain an understanding of the fundamental concepts of algorithms, the very heart of computer science. It introduces the basic data structures and programming techniques often used in efficient algorithms. Covers use of lists, push-down stacks, queues, trees, and graphs. Later chapters go into sorting, searching and graphing algorithms, the string-matching algorithms, and the Schonhage-Strassen integer-multiplication algorithm. Provides numerous graded exercises at the end of each chapter.


0201000296B04062001

The Design and Analysis of Computer Algorithms

The modern science of networks has brought significant advances to our understanding of complex systems. One of the most relevant features of graphs representing real systems is community structure, or clustering, i. e. the organization of vertices in clusters, with many edges joining vertices of the same cluster and comparatively few edges joining vertices of different clusters. Such clusters, or communities, can be considered as fairly independent compartments of a graph, playing a similar role like, e. g., the tissues or the organs in the human body. Detecting communities is of great importance in sociology, biology and computer science, disciplines where systems are often represented as graphs. This problem is very hard and not yet satisfactorily solved, despite the huge effort of a large interdisciplinary community of scientists working on it over the past few years. We will attempt a thorough exposition of the topic, from the definition of the main elements of the problem, to the presentation of most methods developed, with a special focus on techniques designed by statistical physicists, from the discussion of crucial issues like the significance of clustering and how methods should be tested and compared against each other, to the description of applications to real networks.

/pdf/community-detection-in-graphs-34rogm3r6p.pdf

Community detection in graphs

Many web-based systems have a tiered application architecture, in which a request needs to transverse all the tiers before finishing its processing. One of the most important QoS metrics for these applications is the expected response time for the user. Since the expected response time in any tier depends upon the number of servers allocated to this tier, and a request’s total response time is the sum of the response times at all the tiers, many different configurations (number of servers allocated to each tier) can satisfy the expected response time requirement. Naturally, one would like to find the configuration to minimize the total system cost while satisfying the total response time requirement. This is modeled as a non-linear optimization problem using an open-queuing network model of response time, which we call the server allocation problem for tiered systems (SAPTS).

In this paper we study the computational complexity of SAPTS and design efficient algorithms to solve it. For a variable number of tiers, we show that the decision problem of SAPTS is NP-complete. Then we design a simple two-approximation algorithm and a fully polynomial time approximation scheme (FPTAS). If the number of tiers is a constant, we show that SAPTS is polynomial-time solvable. Furthermore, we design a fast polynomial-time exact algorithm to solve for the important two-tier case. Most of our results extend to the general case where each tier has an arbitrary response time function.

Server allocation algorithms for tiered systems

Finding a feasible flow in a strongly connected network

The multigrid method is a general and powerful means of accelerating the convergence of discrete iterative methods for solving partial differential equations PDEs and similar problems. The adaptation of the multigrid method to un structured meshes is important in solving problems with complex geometries. Such problems lie on the forefront of many scientific and engineering fields. Unfortunately, multi grid schemes on unstructured meshes require signifi cantly more preprocessing than on structured meshes. In fact, preprocessing can be a major part of the solution task and, for many applications, must be executed repeatedly. In addition, the large computational requirements of real istic PDEs, accurately discretized on unstructured meshes, make such computations candidates for parallel or distributed processing. This adds problem partitioning as a preprocessing task. We propose and examine experi mentally an automatic and unified strategy to perform several unstructured multigrid preprocessing tasks. Our strategy is based on dominating sets in the unstructured meshes. We also suggest several alternative related strategies. Our experiments evaluate the performance of two preprocessing tasks: coarse-mesh generation and domain partitioning. The experiments suggest that our preprocessing strategy produces high-quality meshes that give good multigrid performance. Our strategy also pro duces domain partitions that are reasonably load bal anced with relatively small edge cuts. Overall, we conclude that simple, integrated algorithmic strategies and data structures can make tedious preprocessing tasks more efficient and more automated-a necessary step toward the practical application of unstructured multigrid methods.

Toward Efficient Unstructured Multigrid Preprocessing

: The classical maximum flow problem sometimes occurs in settings in which the capacities are not fixed but are functions of a single parameters, and the goal is to find the value of the parameter such that the corresponding maximum flow or minimum cut satisfies some side condition. Finding the desired parameter value requires solving a sequence of related maximum flow problems. We shoe that the recent maximum flow algorithm of Goldberg and Tarjan can be extended to solve an important class of such parametric maximum flow problems, at the cost of only a constant factor in its worst case time bound. Faster algorithms for a variety of combinational optimization problems follow from our result. Keywords: Algorithms; Data structures; Graphs; Maximum glow; Network flows; Networks.

A Fast Parametric Maximum Flow Algorithm. Revision

The quest for efficiency in computational methods yields not only fast algorithms, but also insights that lead to elegant, simple, and general problem-solving methods.

Robert E. Tarjan

Papers

Server allocation algorithms for tiered systems

Finding a feasible flow in a strongly connected network

Toward Efficient Unstructured Multigrid Preprocessing

A Fast Parametric Maximum Flow Algorithm. Revision

Algorithm design