Core AI Research Articles

Redesigning an AI bill of rights with/for young people: Principles for exploring AI ethics with middle and high school students

“AI ethics” has become a popular term for signaling concerns with the development and application of artificial intelligence, and a hot-button issue in both academic and public discourse. The current work tends to center adult-oriented, adult-generated, and issues-based conceptions of ethics. This paper reports on an alternative model for “doing ethics” via co-design with young people, highlighting the need for students to engage in AI ethics work, especially as AI is increasingly used in their education and social lives. In particular, we focus on developing a rights-based approach, using the White House Blueprint for an AI Bill of Rights as a starting point for redesigning core AI ethics principles and their conveyance to be accessible and relevant to young people, in collaboration with young people. In this manuscript, we describe our empirical work in a series of three studies on AI ethics teaching and learning with middle and high school students, which culminated in redesigning pieces of the AI Bill of Rights. We present lessons learned, implications for responsible AI in education from student perspectives, and conclude with our suggested principles for integrating student insight and perspectives into the ethics frameworks that guide AI applications.

(Invited, Digital Presentation) Overview of Current and Future 3D Heterogeneous Integration Architectures

High performance computing architectures such as high core count servers and AI accelerators have increasing demands on performance and energy efficiency. This requires power efficient and high bandwidth integration of multiple functions such as processing, memory and high speed circuits in a single package. Current trend is to separate these functions into chiplets with optimized technology then integrating them using advanced packaging technologies that target ease of integration, low overhead and maximum flexibility. In this talk, we will provide an overview of current and future 3D packaging technologies. We will start with solder silicon interposers which enable die stacking to support improved logic density/mm2 of package area and cover their design considerations and applications. Next, we will cover local silicon interconnect architectures which bridge the flexibility gap between 2.5D architectures and 3D interposer architectures. Afterwards, we will discuss the next generation architectures which are based on hybrid bonding. These architectures offer much higher interconnect density up to >100,000 connections/mm2 and enable new types of integration. Finally, we will discuss some future directions to continue to support performance and bandwidth scaling.

수학과 인공지능(AI) 핵심 개념과 <인공지능 수학> 내용 체계 분석

In this study, ‘data collection’, ‘data expression’, ‘data analysis, and ’optimization and decision-making’ were selected as the core AI concepts to be dealt with in the mathematics for AI education. Based on this, the degree of reflection of AI core concepts was investigated and analyzed compared to the mathematical core concepts and content of each area of the elective course. In addition, the appropriateness of the content of <AI Mathematics> was examined with a focus on core concepts and related learning contents. The results provided some suggestions for answering the following four critical questions. First, How to set the learning path for <AI Mathematics>? Second, is it necessary to discuss the redefinition of the nature of <AI Mathematics>? Third, is it appropriate to select core concepts and terms for <Artificial Intelligence Mathematics>? Last, is it appropriate to present the relevant learning contents of the content system of <AI Mathematics>?

The Forgotten Third: A Comparison of China, US, and the European Union’s AI Development

In the 21st century, great power competition dominates the field of international relations. Much has been written about the US and China rivalry for technological dominance, specifically in artificial intelligence. But these analyses are missing one essential player: Europe. I ask will China use its advancements in artificial intelligence to overtake the United States as a superpower, disrupting the US hegemon, just as the United States once did in a post-cold war era with the USSR. Europe is developing its own strategy and capabilities to rival those of the US and China. I use a cross-country qualitative case study method to examine advancements in artificial intelligence across the US, China, and Europe, specifically France and Germany. To determine each states’ leadership and capabilities, I compare them across their AI dreams, hardware, research, and ecosystem. In this comparison I find that whilst China’s numbers outcompete the US and Europe in total output, there are multiple criterium, notably in top tier development, where there is still a significant gap China needs to close between its rivals. Thus, providing an opportunity for Europe, specifically France and Germany, to develop and lead certain criterium regarding core AI development. This paper contributes to existing scholarship on artificial intelligence and US-China relations by adding the European dimension.

AI fun matters

Across: 1 ) Delivered by train. 7 ) Give a false impression. 12 ) With no parts missing. 13 ) Unlimited warranty? 14 ) Join an ideology. 15 ) Core AI technology. 16 ) Evolve as the Pisa Tower. 17 ) Rewarded for work. 19 ) Girl mentioned in absence. 20 ) Apple computer. 21 ) Queen goddess in Mt. Olympus. 22 ) Unpleasant look. 23 ) Catching with a cruel, inhumane trap. 25 ) Puts together a book or a thesis. 26 ) Suppliers of a product or service. 28 ) Living plagues. 31 ) Passes... into the fourth dimension. 35) __ = Distance/Time. 36 ) Naval military prison. 37 ) Set an upper bound. 38 ) Deborah __, Scotish journalist. 39 ) Competitive speed-wise. 40 ) Two in a binary logarithm. 41 ) Old computer bus design. 43 ) Spanish airline. 45 ) Biggest prime digits. 46 ) Sport from Electronic Arts' FIFA series. 47 ) Fashion movement. 48 ) Adopters of a greedy approach.

Shaping the Law of AI: Transatlantic Perspectives

Shaping the Law of AI: Transatlantic Perspectives

Compressed Path Databases with Ordered Wildcard Substitutions

Compressed path databases (CPDs) are a state-of-the-art approach to path planning, a core AI problem. In the Grid-based Path Planning Competition, the CPD-based SRC path planning system was the fastest competitor with respect to both computing full optimal paths and computing the first moves of an optimal path. However, on large maps, CPDs can require a significant amount of memory, which can be a serious practical bottleneck. We present an approach that significantly reduces the size of a CPD. Our approach replaces part of the data encoded in a CPD with wildcards ("don’t care" symbols), maintaining the ability to compute optimal paths for all pairs of nodes of an undirected graph. We show that using wildcards in a way that maximizes the memory savings is NP-hard. We consider heuristics that achieve a good performance in practice. We implement our ideas on top of SRC and provide a detailed empirical analysis. Average memory savings can reach a factor of 2. Our first-k-moves lag (i.e., the time before knowing the first k optimal forward moves) increases, but it can be kept within competitive values. The speed of computing full optimal paths improves slightly.

Often, It's Not about the AI

Narrowly focused task and domain specific AI has been applied successfully for more than twenty five years, and has produced immense value in industry and government. It doesn’t lead directly to artificial general intelligence (AGI), but it does have real problem solving value. It is useful to note that many of the reasons why some otherwise meritorious AI applications fail have nothing to do with the AI per se, but rather, with systems engineering and organizational issues. For example: the domain expert is pulled out to work on more critical projects; the application champion rotates out of his/her position; or the sponsor changes priorities. A system may not make it past an initial pilot test for logistical vs. substantive technical reasons. Some embedded AI systems may work well for years on a software platform that is orphaned and porting it would be prohibitively expensive. A system may work well in a pilot test, but it might not scale for huge numbers of users without extensive performance optimization. The core AI system may be great but the user interface could be suboptimal. The delivered application system might work well, but it could be hard to maintain internally. The system may work according to the sponsor’s requirements, but it might not be applied to the part of the problem that delivers the largest economic results; or the system might not produce enough visible organizational benefits to protect it in subsequent budget battles. Alternatively, the documented results may be quite strong, but may not be communicated effectively across organizational boundaries. All software projects are vulnerable to one or more of these problems. The fact that some software projects have a relatively small percentage of their total code in embedded AI methods doesn’t make them an exception. However, knowing about these potential problems could help AI project teams to be proactive about avoiding them whenever possible.

Cambrian intelligence: the early history of the new AI

Until the mid-1980s, AI researchers assumed that an intelligent system doing high-level reasoning was necessary for the coupling of perception and action. In this traditional model, cognition mediates between perception and plans of action. Realizing that this core AI, as it was known, was illusory, Rodney A. Brooks turned the field of AI on its head by introducing the behavior-based approach to robotics. The cornerstone of behavior-based robotics is the realization that the coupling of perception and action gives rise to all the power of intelligence and that cognition is only in the eye of an observer. Behavior-based robotics has been the basis of successful applications in entertainment, service industries, agriculture, mining, and the home. It has given rise to both autonomous mobile robots and more recent humanoid robots such as Brooks' Cog. This book represents Brooks' initial formulation of and contributions to the development of the behavior-based approach to robotics. It presents all of the key philosophical and technical ideas that put this bottom-up approach at the forefront of current research in not only AI but all of cognitive science.

Intelligence without robots: a reply to Brooks

In his recent papers, entitled Intelligence without Representation and Intelligence without Reason, Brooks argues for mobile robots as the foundation of AI research. This article argues that even if we seek to investigate complete agents in real-world environments, robotics is neither necessary nor sufficient as a basis for AI research. The article proposes real-world software environments, such as operating systems or databases, as a complementary substrate for intelligent-agent research and considers the relative advantages of software environments as test beds for AI. First, the cost, effort, and expertise necessary to develop and systematically experiment with software artifacts are relatively low. Second, software environments circumvent many thorny but peripheral research issues that are inescapable in physical environments. Brooks's mobile robots tug AI toward a bottom-up focus in which the mechanics of perception and mobility mingle inextricably with or even supersede core AI research. In contrast, the softbots (software robots) I advocate facilitate the study of classical AI problems in real-world (albeit, software) domains. For example, the UNIX softbot under development at the University of Washington has led us to investigate planning with incomplete information, interleaving planning and execution, and a host of related high-level issues.

Foundations of AI: The big issues

The objective of research in the foundations of AI is to explore such basic questions as: What is a theory in AI? What are the most abstract assumptions underlying the competing visions of intelligence? What are the basic arguments for and against each assumption? In this essay I discuss five foundational issues: (1) Core AI is the study of conceptualization and should begin with knowledge level theories. (2) Cognition can be studied as a disembodied process without solving the symbol grounding problem. (3) Cognition is nicely described in propositional terms. (4) We can study cognition separately from learning. (5) There is a single architecture underlying virtually all cognition. I explain what each of these implies and present arguments from both outside and inside AI why each has been seen as right or wrong.