Computational Context: The Value, Theory and Application of Context with AI
暫譯: 計算背景：AI 中背景的價值、理論與應用
William F. Lawless , Ranjeev Mittu , Donald Sofge
商品描述
This volume addresses context from three comprehensive perspectives: first, its importance, the issues surrounding context, and its value in the laboratory and the field; second, the theory guiding the AI used to model its context; and third, its applications in the field (e.g., decision-making). This breadth poses a challenge. The book analyzes how the environment (context) influences human perception, cognition and action. While current books approach context narrowly, the major contribution of this book is to provide an in-depth review over a broad range of topics for a computational context no matter its breadth. The volume outlines numerous strategies and techniques from world-class scientists who have adapted their research to solve different problems with AI, in difficult environments and complex domains to address the many computational challenges posed by context.
Context can be clear, uncertain or an illusion. Clear contexts: A father praising his child; a trip to the post office to buy stamps; a policewoman asking for identification. Uncertain contexts: A sneak attack; a surprise witness in a courtroom; a shout of "Fire Fire " Contexts as illusion: Humans fall prey to illusions that machines do not (Adelson's checkerboard illusion versus a photometer). Determining context is not easy when disagreement exists, interpretations vary, or uncertainty reigns. Physicists like Einstein (relativity), Bekenstein (holographs) and Rovelli (universe) have written that reality is not what we commonly believe. Even outside of awareness, individuals act differently whether alone or in teams.
Can computational context with AI adapt to clear and uncertain contexts, to change over time, and to individuals, machines or robots as well as to teams? If a program automatically "knows" the context that improves performance or decisions, does it matter whether context is clear, uncertain or illusory? Written and edited by world class leaders from across the field of autonomous systems research, this volume carefully considers the computational systems being constructed to determine context for individual agents or teams, the challenges they face, and the advances they expect for the science of context.
商品描述(中文翻譯)
這本書從三個全面的角度探討了上下文：首先是其重要性、圍繞上下文的問題以及其在實驗室和現場的價值；其次是指導用於建模上下文的人工智慧理論；第三是其在現場的應用（例如，決策）。這種廣度帶來了挑戰。書中分析了環境（上下文）如何影響人類的感知、認知和行動。雖然目前的書籍對上下文的探討較為狹隘，但本書的主要貢獻在於提供對計算上下文的廣泛主題的深入回顧，無論其範圍如何。本書概述了來自世界級科學家的眾多策略和技術，他們已經調整自己的研究以解決不同問題，並利用人工智慧在困難的環境和複雜的領域中應對上下文所帶來的許多計算挑戰。
上下文可以是清晰的、不確定的或是一種幻覺。清晰的上下文：一位父親讚美他的孩子；去郵局買郵票的旅行；一位女警要求出示身份證明。不確定的上下文：突襲；法庭上的驚喜證人；一聲「火災！火災！」作為幻覺的上下文：人類會受到機器不會受到的幻覺影響（阿德爾森的棋盤幻覺與光度計）。當存在分歧、解釋各異或不確定性主導時，確定上下文並不容易。像愛因斯坦（相對論）、貝肯斯坦（全息圖）和羅維利（宇宙）等物理學家曾寫道，現實並不是我們通常所相信的樣子。即使在不自覺的情況下，個體在獨自或團隊中行動的方式也會有所不同。
計算上下文與人工智慧能否適應清晰和不確定的上下文，隨著時間和個體、機器或機器人以及團隊的變化而變化？如果一個程序自動「知道」能改善性能或決策的上下文，那麼上下文是清晰的、不確定的還是虛幻的，這是否重要？本書由來自自主系統研究領域的世界級領導者撰寫和編輯，仔細考慮了正在構建的計算系統，以確定個體代理或團隊的上下文、他們面臨的挑戰以及他們對上下文科學的期望進展。
作者簡介
William Lawless, as an engineer, in 1983, Lawless blew the whistle on Department of Energy's mismanagement of radioactive wastes. For his PhD, he studied the causes of mistakes by organizations with world-class scientists and engineers. Afterwards, DOE invited him onto its citizen advisory board at its Savannah River Site where he co-authored numerous recommendations on the site's clean-up. In his research on mathematical metrics for teams, he has published two co-edited books on AI, and over 200 articles, book chapters and peer-reviewed proceedings. He has co-organized eight AAAI symposia at Stanford (e.g., in 2018: Artificial Intelligence for the Internet of Everything).
Ranjeev Mittu, is a Branch Head for the Information Management and Decision Architectures Branch within the Information Technology Division at the U.S. Naval Research Laboratory. He is the Section Head of Intelligent Decision Support Section which develops novel decision support systems through applying technologies from the AI, multi-agent systems and web services. He brings a strong background in transitioning R&D solutions to the operational community, demonstrated through his current sponsors including DARPA, OSD/NII, NSA, USTRANSCOM and ONR. He has authored 2 books, 5 book chapters, and numerous conference publications. He has an MS in Electrical Engineering from Johns Hopkins University.
Donald (Don) Sofge is a Computer Scientist and Roboticist at the U.S. Naval Research Laboratory (NRL) with 30 years of experience in Artificial Intelligence and Control Systems R&D. He has served as PI/Co-PI on dozens of federally funded R&D programs and has authored/co-authored approximately 110 peer-reviewed publications, including several edited books, many journal articles, and several conference proceedings. Don leads the Distributed Autonomous Systems Group at NRL where he develops nature-inspired computing solutions to challenging problems in sensing, artificial intelligence, and control of autonomous robotic systems. His current research focuses on control of autonomous teams or swarms of robotic systems.
作者簡介(中文翻譯)
威廉·勞萊斯（William Lawless）是一位工程師，於1983年揭發了美國能源部在放射性廢物管理上的失職。為了取得博士學位，他研究了世界級科學家和工程師所屬組織犯錯的原因。隨後，能源部邀請他加入其位於薩凡納河地區的公民諮詢委員會，他共同撰寫了多項有關該地區清理工作的建議。在他對團隊數學指標的研究中，他共同編輯了兩本有關人工智慧的書籍，並發表了超過200篇文章、書章和經過同行評審的會議論文。他在史丹佛大學共同組織了八場美國人工智慧協會（AAAI）研討會（例如，2018年：為萬物互聯的人工智慧）。
蘭吉夫·米圖（Ranjeev Mittu）是美國海軍研究實驗室（U.S. Naval Research Laboratory）資訊科技部門資訊管理與決策架構分部的分部主管。他是智能決策支援部門的部門主管，該部門通過應用人工智慧、多代理系統和網路服務的技術來開發新穎的決策支援系統。他在將研發解決方案轉移到操作社群方面擁有堅實的背景，這一點在他目前的贊助商中得到了證明，包括DARPA、OSD/NII、NSA、USTRANSCOM和ONR。他著有2本書、5章書籍以及多篇會議出版物。他擁有約翰霍普金斯大學的電機工程碩士學位。
唐納德（唐）索夫格（Donald (Don) Sofge）是美國海軍研究實驗室（NRL）的計算機科學家和機器人專家，擁有30年的人工智慧和控制系統研發經驗。他曾擔任數十個聯邦資助的研發計畫的首席研究員或共同首席研究員，並撰寫或共同撰寫了約110篇經過同行評審的出版物，包括幾本編輯書籍、許多期刊文章和幾篇會議論文。唐在NRL領導分散式自主系統小組，開發自然啟發的計算解決方案，以應對感測、人工智慧和自主機器人系統控制等挑戰性問題。他目前的研究重點是控制自主團隊或機器人系統的群體。
