Robert Ross

Jürgen Dix

V. S. Subrahmanian

Sarit Kraus

Thomas Eiter

Piero Bonatti

Fatma Ozcan

After a discussion of the theory of software agents, this book presents IMPACT(Interactive Maryland Platform for Agents Collaborating Together), an experimental agentinfrastructure that translates formal theories of agency into a functional multiagent system thatcan extend legacy software code and application-specific or legacy data structures.

Heterogeneous Agent Systems

Until recently, databases contained easily indexed numbers and text. Today, in the age of powerful, graphically based computers, and the world wide web, databases are likely to contain a much greater variety of data forms, including images, sound, video clips, and even handwritten documents. When multimedia databases are the norm, traditional methods of working with databases no longer apply. How do you query a video library, or an image database containing x-rays, or sounds in an audio database? Principles of Multimedia Database Systems explains how to work with these new multimedia data forms. It is the first comprehensive treatment of the skills and techniques required to build, maintain, and query multimedia databases. This book presents the mix of techniques necessary for working with multimedia databases, including synthetic solutions for the design and deployment of multimedia database systems. Because rapid technological developments are constantly changing the landscape of multimedia databases, the book teaches basic theoretical principles applicable to any database. * Covers the major issues of multimedia database design, with a strong focus on distributed multimedia databases. * Discusses important topics including how to organize the vast data types, storage and retrieval, and creation and delivery of multimedia presentations. * Organized around the lively scenario of a crime-fighting database that evolves as new concepts are introduced. * Includes numerous exercises and suggestions for programming projects. * Additional materials on the web include updates, on-line supplements, and links to downloadable software.

Principles of Multimedia Database Systems

Judee K. Burgoon

Norah E. Dunbar

This book analyzes the multimodal verbal and nonverbal behavior of humans in both an artificial game, based on the well-known Mafia and Resistance games, as well as selected other settings. This book develops statistical results linking different types of facial expressions (e.g. smile, pursed lips, raised eyebrows), vocal features (e.g., pitch, loudness) and linguistic features (e.g., dominant language, turn length) with both unary behaviors (e.g. is person X lying?) to binary behaviors (Is person X dominant compared to person Y? Does X trust Y? Does X like Y?). In addition, this book describes machine learning and computer vision-based algorithms that can be used to predict deception, as well as the visual focus of attention of people during discussions that can be linked to many binary behaviors. It is written by a multidisciplinary team of both social scientists and computer scientists. Meetings are at the very heart of human activity. Whether you are involved in a business meeting or in a diplomatic negotiation, such an event has multiple actors, some cooperative and some adversarial. Some actors may be deceptive, others may have complex relationships with others in the group. This book consists of a set of 11 chapters that describe the factors that link human behavior in group settings and attitudes to facial and voice characteristics. Researchers working in social sciences (communication, psychology, cognitive science) with an interest in studying the link between human interpersonal behavior and facial/speech/linguistic characteristics will be interested in this book. Computer scientists, who are interested in developing machine learning and deep learning based models of human behavior in group settings will also be interested in purchasing this book.

Detecting Trust and Deception in Group Interaction

James F. Brown

Chiara Pulice

Jacob Bonen-Clark

This is the first study of Boko Haram that brings advanced data-driven, machine learning models to both learn models capable of predicting a wide range of attacks carried out by Boko Haram, as well as develop data-driven policies to shape Boko Haram’s behavior and reduce attacks by them. This book also identifies conditions that predict sexual violence, suicide bombings and attempted bombings, abduction, arson, looting, and targeting of government officials and security installations. After reducing Boko Haram’s history to a spreadsheet containing monthly information about different types of attacks and different circumstances prevailing over a 9 year period, this book introduces Temporal Probabilistic (TP) rules that can be automatically learned from data and are easy to explain to policy makers and security experts. This book additionally reports on over 1 year of forecasts made using the model in order to validate predictive accuracy. It also introduces a policy computation method to rein in Boko Haram’s attacks. Applied machine learning researchers, machine learning experts and predictive modeling experts agree that this book is a valuable learning asset. Counter-terrorism experts, national and international security experts, public policy experts and Africa experts will also agree this book is a valuable learning tool.

A Machine Learning Based Model of Boko Haram

V. S. Subrahmanian

About

Books

Join the club