AAAI-10 / IAAI-10 Invited Speakers
Intelligent Interaction with the Real World
Leslie Pack Kaelbling (Massachusetts Institute of Technology)
Since the inception of the field, one of the visions of artificial intelligence has been robust, intelligent, general-purpose robots that interact with the real world. We have made useful progress in that direction, but there is still a long way to go. I will characterize one view of how we might achieve this goal, describe some intermediate results, and characterize important technical and methodological problems that must be solved to make that vision real.
Leslie Pack Kaelbling is the Ellen Swallow Richards Professor of Computer Science and Engineering at the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology. She has made research contributions to decision-making under uncertainty, learning, and sensing with applications to robotics, with a particular focus on reinforcement learning and planning in partially observable domains.
She holds an A.B in philosphy and a Ph.D. in computer science from Stanford University, and has previously held positions at SRI International, Teleos Research, and Brown University. She is the recipient of the US NSF Presidential Faculty Fellowship, the IJCAI Computers and Thought Award, and several teaching prizes, and is a fellow of the AAAI. She was the founder and editor-in-chief of the Journal of Machine Learning Research.
AAAI-10 Invited Talk
Challenges for AI in Computational Sustainability
Carla P. Gomes (Cornell University)
Computational Sustainability is a new interdisciplinary research field with the overall goal of developing computational models, methods, and tools to help manage the balance between environmental, economic, and societal needs for a sustainable future. In this talk I will provide an overview of Computational Sustainability, with examples ranging from wildlife conservation and biodiversity, to poverty mitigation, to large-scale deployment and management of renewable energy sources. I will highlight overarching computational challenges for AI at the intersection of constraint reasoning, optimization, machine learning, and dynamical systems. Finally I will discuss the need for a new approach that views computational sustainability problems as “natural” phenomena, amenable to a scientific methodology, in which principled experimentation, to explore problem parameter spaces and hidden problem structure, plays as prominent a role as formal analysis.
Carla Gomes is an associate professor of computer science at Cornell University, with appointments in the departments of Computer Science, Information Science, and Applied Economics and Management. Her research spans the full range of theory to applications. A central theme of her research is large scale constraint reasoning and optimization, synthesizing concepts from constraint reasoning, mathematical programming, and machine learning. Gomes is currently pursuing the new research area of computational sustainability. Gomes is the lead principal investigator of an NSF Expeditions in Computing on Computational Sustainability and the director of the newly established Institute for Computational Sustainability at Cornell University. Gomes is a Fellow of the Association for the Advancement of Artificial Intelligence.
AAAI-10 Invited Talk
Constraint Programming and Artificial Intelligence: Challenges, Applications and Opportunities
Barry O’Sullivan (University College Cork)
Constraint programming is a powerful tool for modeling and solving complex optimization problems. It is widely used to support industrial decision-making, as well as being used as a component in various intelligent systems. Constraint programming has its origins in constraint satisfaction and logic programming from the field of artificial intelligence, and mathematical programming from the field of operations research. In this talk I will: identify some of the challenges facing the field; present some exciting applications of constraint programming; and give a view of where opportunities lie for the future from the perspectives of both science and application.
Barry O’Sullivan is a senior lecturer in constraints at the Department of Computer Science at University College Cork in Ireland. He is the associate director of the Cork Constraint Computation Centre, Science Foundation Ireland principal investigator, president of the Association for Constraint Programming, chairman of the Artificial Intelligence Association of Ireland, coordinator of the European Research Consortium for Informatics and Mathematics (ERCIM) Working Group on Constraints, and a member of the Executive Council of the Management Science Society of Ireland. His research interests include real-world applications of artificial intelligence, constraint programming and optimization technologies.
AAAI-10 Invited Talk
Incentive Engineering in the Internet Age
David C. Parkes (Harvard University)
Mechanism design provides a formalism within which to understand the possible and the impossible when designing multi-agent systems with private information and rational agents. In introducing computational considerations, we have gained some understanding of how to reconcile new tensions that arise. Today, we see a thirst for practical, engineered incentive mechanisms to deploy across the myriad of multiuser systems enabled by the Internet. I will highlight some of the new challenges that this presents, in moving from isolated events to continual processes, from simple models to complex, multifaceted agent models, and in enabling new kinds of computational and coordination processes.
David C. Parkes is Gordon McKay Professor of Computer Science at Harvard University and has a Ph.D. degree in computer and information science from the University of Pennsylvania and an M.Eng. (First class) in engineering and computing science from The University of Oxford. A recipient of the NSF CAREER Award, the Alfred P. Sloan Fellowship, and Harvard’s Roslyn Abramson Award for Teaching, Parkes serves as an editor of Games and Economic Behavior and on the editorial boards of JAAMAS and the INFORMS Journal of Computing, is general chair for ACM EC’10 and was program chair for ACM EC’07 and AAMAS’08.
AAAI-10 Invited Talk
Systems with General Intelligence: A New Perspective
Michael Thielscher (The University of New South Wales)
General intelligence enables agents and robots to adapt automatically to previously unknown situations. This new level of autonomy has recently been achieved in Computer Game Playing with the advent of programs that understand the rules of new games and learn to play them well without human intervention. This talk will advocate agents and robots with general intelligence as a Grand Challenge, aiming at the next generation of AI systems that can be told what we expect from them and that learn to carry out radically different tasks in radically new environments. It will be argued that this calls for a new research focus for AI with the goal to integrate different, successful AI methodologies.
Michael Thielscher is ARC Future Fellow and professor at The University of New South Wales. He received his postgraduate diploma in 1992 and his Ph.D. in computer science in 1994, both with distinction, from Darmstadt University. He then joined Dresden University where he was associate professor before he moved to his present position in Sydney in 2010. His habilitation thesis was honored with the Award for Research Excellence by the alumni of Darmstadt University in 1997, and in 2009 he won a Future Fellowship Award from the Australian Research Council. He has made several fundamental contributions to knowledge representation for agents, belief revision, and general game playing. He is author of over 100 refereed papers, four books, and FLUXPLAYER, which was crowned the world champion at the 2006 AAAI General Game Playing Competition.
Robert S. Engelmore Memorial Award Lecture
Cancer: A Computational Disease that AI Can Cure
Jay M. Tenenbaum (CollabRx Inc.)
Cancer results from finite genomic mutations that biotechnology can easily list, and that we can mostly understand and reason about in terms of the underlying biochemistry. Tragically, the scientific and medical communities are searching for cures using an incredibly inefficient non-adaptive strategy, where the costs of experiments are measured in lives, as well as money, and where we capture only a small portion of the genomics and outcomes data, i.e., in clinical trials. Inspired by my career experiences as an AI researcher, Internet entrepreneur and cancer survivor, I am attempting to redress this situation through Cancer Commons, a “rapid learning” community of patients, physicians and researchers. Our goal is to cure cancer by collecting the genomic and response data from thousands of adaptively-planned individual treatment experiments, integrating the resulting sparse fragments of evidence to infer the true causal mechanisms of tumors and drugs, and generalizing the resulting knowledge so that it can be applied to new cases. Each patient is treated in accord with the best available knowledge, and that knowledge is continually updated to benefit the next patient. Hopefully, this adaptive approach will efficiently climb the hill to find cures for cancer, one patient at a time.
Jay M. Tenenbaum was educated at the Massachusetts Institute of Technology and Stanford University in the 1960s. He spent the 1970s doing artificial intelligence research at SRI, the 1980s managing computer science research for Schlumberger, and the 1990s pioneering Internet commerce. He’s currently focused on using the AI and the web to transform medicine.
IAAI-10 Invited Talk
A New Paradigm of Geriatric Care Empowered by Applied AI
Majd Alwan (Center for Aging Services Technologies)
Advances in sensor, communication, artificial intelligence technologies and data processing, coupled the increasing processing power, is causing a shift in the way we care for the elderly. This paper presents a new paradigm for geriatric care based on monitoring and assisting older adults in their own living settings. AI techniques could be applied to mine health and activity data collected in the home to detect indicators of early disease onset, deterioration or improvement, inform providers and allow the delivery of care services, including assistance and support. Examples of monitoring and assistive systems that apply AI techniques are discussed. The approach has significant value to older adults as well as caregivers, and allows care providers to extend services into the community.
Majd Alwan is the director of the Center for Aging Services Technologies (CAST). He is responsible for creating and leading a network of technology companies, providers and research institutions focused on technology solutions for an aging society. Prior to joining CAST, Majd served as an assistant professor and the director of the robotics and eldercare technologies program at the University of Virginia’s Medical Automation Research Center. His research interests included passive functional and health assessment, biomedical instrumentation, as well as eldercare and assistive technologies.
Alwan received his Ph.D. in intelligent robotics from Imperial College of Science, Technology and Medicine, University of London, a Master’s of Science degree in control engineering with distinction from Bradford University and a bachelor’s degree in electrical engineering from Damascus University. He is a Senior Member of the IEEE.
IAAI-10 Invited Talk
Species of Mind
Vernor Vinge (San Diego State University)
More than any other animal, we humans invent ways to outsource cognitive function. We’ve been doing this for a long time. For instance, writing is an outsourcing of memory; money is a scalar that allows the comparison of vastly different objects.
During the last century, the outsourcing process has become more diverse and intense. The range of our recent activities is leading toward a number of “different kinds” of superhuman intelligence.
In this presentation, I’ll discuss several different paths to superintelligence, their relative power, the transformations they might create, and how humans might deal with them.
Vernor Vinge is a science-fiction writer, winner of five Hugo Awards. He holds a PhD (mathematics) from University of California, San Diego and taught mathematics and computer science at San Diego State University from 1972 to 2000. In 1982, at a panel for AAAI-82, he proposed that technology would accelerate the evolution of intelligence, leading to a kind of “singularity” beyond which merely human extrapolation was essentially impossible. In the 1980s and 1990s, he elaborated on this theme in both his science fiction and nonfiction writing.