Promotions - Choice's Outstanding Academic Titles 2020 -

Mitchell (Portland State Univ.) has authored one of the most readable books on artificial intelligence to have appeared in a long time. How so? First and most important, Mitchell understands and has mastered the craft of writing. She is telling us a story, and yet, though it reads like a novel, the story is about the state of the art in the contemporary field of AI. The text is composed of 16 chapters organized into 5 parts. Part 1, "Background," establishes the foundation, including Mitchell's motivation for writing, concluding with the third chapter, entitled "AI Spring." The three or four chapters that make up each subsequent part of the book transport readers to each of the later seasons of AI. Part 2, "Looking and Seeing," introduces us to machine learning. Part 3, "Learning to Play," introduces robots. Part 4, "Artificial Intelligence Meets Natural Language," while seemingly self-explanatory, reveals how natural language processing crossed over into the territory of "the last 10 percent" around 2012. Part 5, "The Barrier of Meaning," offers a new appreciation of the real current state, as well as the future, of AI. The book concludes with an exhaustive index and a section of chapter notes.Summing Up: Highly recommended. Lower-division undergraduates through faculty and professionals.

Blockchain is an internet technology that maintains history for digital objects. This history is distributed and encrypted, securing the provenance of the digital object while protecting privacy for those who interacted with the digital object. Blockchain is a disruptive technology; for example, it allows Bitcoin to provide a secure alternative currency system that operates entirely outside government or banking regulation and protects the anonymity of its users. This confounds and undermines the authority of traditional financial, monetary, and policing institutions and at the same time introduces a new kind of trust foundation for society--or perhaps an alternative to trust based on blockchain algorithms. This monograph edited by Ragnedda (Northumbria Univ.) and Destefanis (Brunel Univ.) includes 16 chapters describing a wide range of interdisciplinary possibilities for and challenges related to using blockchain for social interaction and trust, customer relationship management, dispute resolution and smart contracts, collective action and global-scale governance, news and journalism, intellectual property protection, global distribution of wealth, and the control and capitalization of personal data. The book is effective both in conveying the range of impacts blockchain has already imposed and in suggesting that much more could yet be in store.Summing Up: Highly recommended. Lower- and upper-division undergraduates. Graduate students, faculty, and professionals.

Stewart (Univ. of Iowa) explores continuities and discontinuities inherent in the century-long transition of moving pictures from celluloid film stock to digital media. Arguing that cinema at the current stage is merely "postfilmic" rather than "postcinematic," his text focuses on apparatus reading, or "studious looking"--approaching the moving picture narrative through its technics while insisting that cinema has always been an "engineered illusion." Along with Stanley Cavell and Christian Metz (as, e.g., in Impersonal Enunciation, or the Place of Film, CH, Aug'16, 53-5185), Stewart here makes extensive use of innovative ideas from Jean Epstein's Intelligence of a Machine (first available in English as translated by Wall-Romana in 2016) while trying out his approach on paradigmatic examples drawn chiefly from classic comedy and science fiction films. Whatever the technology, Stewart argues, moving picture narratives are always articulated by specific "cinemachines" whose presence is made sometimes more, sometimes less evident. Though written with erudite wit and humor, Stewart's text is also laden with enough specialized language and wordplay to narrow his audience to those moderately well-versed in the methodology and styles of contemporary media studies, especially as inspired by the technical metaphor, inscribed by Metz, of cinema as articulated speech. For this audience, Cinemachines will likely serve as an essential reference.Summing Up: Essential. Upper-division undergraduates. Graduate students and faculty researchers.

Blum (Toyota Technical Institute, Univ. of Chicago), Hopcroft (Cornell Univ.), and Kannan (Microsoft Research, India) have written a graduate level textbook (here abbreviated as BHK) on the mathematical foundations of data science from a theoretical computer science perspective. Drafts of BHK from 2013 onward can be found online, and BHK has already been used in courses at Cornell, the Bonn-Aachen International Center for Information Technology, and the University of California, Santa Barbara. The authors have "written this book to cover the theory [they] expect to be useful in the next 40 years" with an "increased emphasis on probability, statistics and numerical methods." One plausible measure of BHK's impact is the book's own citation metrics. Semantic Scholar (https://www.semanticscholar.org) reports 81 citations with 42 citations related to background or methods; BHK appears to be on course to becoming influential. Each chapter contains a wealth of exercises and a "Bibliographic Notes" section referencing pertinent source material. BHK is not suited for self-study, however, as there are few solved examples. To address scenarios in which BHK might be used in an undergraduate course, prerequisite theory is helpfully given in the final chapter. Microsoft (https://www.microsoft.com/en-us/research/publication/foundations-of-data-science-2/) hosts nine recorded lectures online, providing a nice preview of what to expect in BHK.Summing Up: Highly recommended. Graduate students and faculty researchers.

Parkeh (Jadavpur Univ., Kolkata, India) offers an elegant expression of the mathematics and MATLAB code for creating 2D and 3D graphics, showing through a succession of steps how increasingly complex graphical objects are created in MATLAB. The journey begins with a simple graph, a line defined only by its endpoints and slope--a linear spline. Next are curved splines, from simple to complex, defined by control points shown on or around the spline, followed by two-dimensional shapes (planes), then three-dimensional objects, including surfaces and projections. Transformations are also included: moving (translating); resizing (scaling); rotating; reflecting; and shearing along the center and along a fixed point. Each stage of this presentation begins with an intuitively graspable image, enabling prior visual understanding of the type of graphical object to be demonstrated. Next, a mathematical framework is provided for building the formula that expresses the object, followed by the step-by-step calculation of the particular object. Finally, relevant MATLAB code is given to mathematically calculate and display the example, with each newly used MATLAB function highlighted, so that by the end of the text a robust graphical vocabulary has been developed. While readers will benefit from some prior exposure to differential calculus and matrices, this book does an excellent job of explaining advanced mathematical concepts.Summing Up: Highly recommended. Upper-division undergraduates. Graduate students and faculty.

This is an excellent resource for scientists and engineers looking to develop a deep natural language understanding system for application in robotics, currently one of the most critical areas of development for artificial intelligence technologies. The interface bandwidth between existing robotics systems and the environment is narrow. Vision-based systems offer the apparent solution. However, improvement in our ability to communicate with robots using natural language would enable a revolutionary bandwidth increase. Yokota (Fukuoka Institute of Technology) has developed a comprehensive approach to natural language understanding based on the assumption that human-level perception relies on omnisensory mental image processing. The mental image directed semantic theory (MIDST) detailed here is an essential contribution to natural language processing research. Yokota's research applies primarily to robotic systems. That said, integrative multimedia understanding (operationalized through a formal language as proposed here) will become foundational to a broader spectrum of applications outside robotics: general multimedia understanding systems. The book is very well written and easy to follow for readers with a good background in artificial intelligence or natural language processing. The considerable number of examples and diagrams included make this volume particularly valuable for students. The examples illustrating machine understanding of Japanese offer an exciting bonus for readers engaged in machine translation or computational linguistics.Summing Up: Highly recommended. Upper-division undergraduates through faculty; professionals.

All online data that users generate is stored because there is a lot of actionable information hidden there. Written posts, comments, images, dialogues, and other forms of online social interaction reveal much about what users think about, including their moods and feelings. It scarcely needs mention that user purchasing preferences and shopping behaviors are also part of the mix. All this data, whether structured or unstructured, may be used to generate machine learning models that can be applied in new contexts, e.g., matching advertising campaigns with the appropriate group of consumers. Such models may also be used for so-called transfer learning tasks, where artificial intelligence (AI) is constructed to broaden the scope of conclusions drawn from known to seemingly unrelated characteristics of system users and use cases. This book dives deep into existing research on user modeling and AI. The text does a great job of explaining how diverse social media data sets and contexts may be used, leading to illustrative discussion about AI algorithms such as classification, sentiment analysis, and recommendation systems. Programmers and systems designers seeking to implement AI systems able to exploit data obtained from social media interactions will benefit from reading this well-written book, assuming they have moderate exposure to AI or machine learning.Summing Up: Highly recommended. All readers.

This inaugural handbook introduces the field of cyberpsychology by elucidating multiple aspects of the online user behavior that has been of popular and academic interest for over 25 years. The editors brought together 60 experts in the field from around the world to pen 34 detailed, well-referenced entries covering eight broad areas of cyberpsychology. These are distributed in the following sections: part 1, "Introduction and Foundations"; part 2, "Technology across the Lifespan"; part 3, "Interaction and Interactivity"; part 4, "Groups and Communities"; part 5, "Social Media"; part 6, "Health and Technology"; part 7, "Gaming"; and part 8, "Cybercrime and Cybersecurity." While the first few entries are foundational, each entry can stand by itself. This reviewer especially enjoyed reading entry 28, "Psychosocial effects of gaming," where the authors departed from the usual textual format of the book by providing an informative and systematic review of the literature. Overall, this handbook provides an all-encompassing, contemporary, and authoritative resource for students and researchers interested in the psychological aspects of how humans and computers interact. The user-friendly ebook version of the handbook is available via Oxford Handbooks Online (CH, Jul'13, 50-5912).Summing Up: Highly recommended. Lower-division undergraduates through faculty.

Smith (Univ. of Toronto) distinguishes here between what he calls reckoning (a product of the current capability of artificial intelligence) and judgment (what sets human intelligence apart). Judgment, he says, is "a form of dispassionate deliberative thought, grounded in ethical commitment and responsible action." While tracing the history of AI from its first "wave," involving mainly reliance on symbolic representation and manipulation of data (sometimes represented as good old fashioned artificial intelligence or GOFAI), to its current phase, realized as machine learning and even deep learning, Smith asserts that these known capabilities are still far from judgment, which would require computers to "know what they are talking about." The argument of this book is that for computers to exhibit general artificial intelligence (GAI) that is comparable to human intelligence, they will need to "give a damn." Smith is interested in deep philosophical questions, beginning with "how" and "why" AI should reach the judgment stage. As stated in the jacket copy, some of his views are controversial. Yet this is an important work exploring questions that should concern readers inside and outside the computer science community. It deserves a wide reading among both audiences.Summing Up: Highly recommended. Upper-division undergraduates through facultly and professionals. General readers.

Brugger (Aarhus Univ., Denmark) and Milligan (Univ. of Waterloo, Ontario) provide a wonderful guide to virtual life since the development of the World Wide Web. This is a monumental tome of 6 parts and 40 chapters, contributed by a total of 48 expert authors. Part 1 ("The Web and Historiography") introduces the web as a primary source. Part 2 ("Theoretical and Methodological Reflections") provides guidance on how to do web history, incorporating both intellectual perspectives and technical approaches, such as network analysis and text mining. Part 3 ("Technical and Structural Dimensions of Web History") provides basic technical information about web protocols (notably Hypertext Transfer Protocol, or HTTP), page-to-page navigation via hyperlinks, search tools, and browser interfaces. Part 4 ("Platforms on the Web") covers Wikipedia, blogging software, and the history of web advertising and archiving. Part 5 ("Web History and Users"), the largest section, has 14 use cases ranging from the architecture of top level domains to preserving the memory of WWI through web archives. Part 6 ("The Roads Ahead") looks at the web's troubled past and hopefully more promising future. An excellent handbook for acquisition by university and local libraries.Summing Up: Highly recommended. All readers.

"Transfer learning" refers to artificial intelligence techniques for machine learning that focus on adapting methods developed for one task and applying them to a new task. Typically, machine learning algorithms are developed with the help of training data suitable for each application. In the transfer learning approach, existing algorithms can potentially be reused, or adapted, for situations presenting a "sparse data" problem. This book offers a comprehensive overview of the field, arguing the case for adaptation as key to mimicking human intelligence. The book is organized in two parts, the first consisting of a thorough methodological review, i.e., types of machine learning: supervised/unsupervised, feature-based, model-based, relation-based, even "lifelong." Part 2 explores a wide range of applications. The intended audience is advanced students, faculty, and researchers. The named "authors" are, in fact, primarily editors. The various chapters were written by graduate students, postdoctoral fellows, and research associates of the senior editor, Yang (Hong Kong Univ. of Science and Technology). The book includes a substantial bibliography documenting copious citations to the literature. There appear to be few other textbooks in this field apart from this unique work. As such, it will be welcomed by libraries supporting strong computer science programs that may have need for a core text in artificial intelligence.Summing Up: Highly recommended. Upper-division undergraduates through faculty; professionals.

Virtual reality (VR) is a hot topic especially among developers responsible for implementing computer graphics technologies. VR is used today by architects, industrial designers, artists, and many others who require immersive visualizations. Digital entertainment is another area of rapid growth. Sherman (Indiana Univ.) provides an excellent resource that encompasses a balanced selection of topics, from history of the VR medium to algorithms, presenting theoretical background as well as code. The title word "gems" signals that the volume is not intended as a textbook per se, to be read in linear fashion. Yet, the chapters are arranged logically into ten thematic sections, such that readers can easily move among different chapters as needed. Section 2 discusses game engines, the foundation of many VR implementations. Next follow contributed chapters on interactive and social aspects of VR: agents, POV cameras, virtual environments, and rendering techniques. The book concludes with two chapters focusing on hardware infrastructure. Many examples throughout depend on employing the Unity game engine as a development platform, with relevant pseudocode examples provided, along with excellent visuals including very informative screenshots of the user interface. This is a highly valuable resource, chiefly for developers with considerable technical background in computer graphics and/or programming.Summing Up: Highly recommended. All readers.