Keynotes : The 4th International Conference on Pervasive Computing (PERVASIVE 2006): Dublin Ireland, May 7-10, 2006

"Off the Beaten Path in Pervasive Computing:
How a Broader Perspective Can Lead to
Unexpected Discoveries and Opportunities"

Joe Marks, Mitsubishi Electric Research Laboratories (MERL)
Cambridge, Massachusetts, USA.
www.merl.com

Joe Marks grew up in Dublin, Ireland, before emigrating to the U.S. to attend college. He holds three degrees from Harvard University. His areas of interest include computer graphics, human-computer interaction, and artificial intelligence. He has worked previously at Bolt Beranek and Newman and at Digital's Cambridge Research Laboratory. He is currently the Director of MERL Research. He is SIGGRAPH 2007 Conference Chair and has was the papers co-chair for Eurographics 2005, the papers chair for SIGGRAPH 2004 and was the the recent past chair of ACM SIGART.

Talk Abstract:

Many of our current projects at MERL are squarely in the pervasive-computing realm, such as: interactive multi-user tabletop displays, operator-identifying automotive controls, digital home networking, UWB-radio location-based services, location-aware RFID tags, semi-passive RFID tags for sensor networks, novel physical and chemical sensors, speech-based information retrieval for mobile devices, and vision-based person/object tracking. But because this work has been or will be presented in forums familiar to this audience, I won’t be talking about these projects today.

Instead, I hope to persuade you that the world of Pervasive Computing is bigger than it seems and that a broader perspective can lead to unexpected discoveries and opportunities. I will use these MERL projects to make my point:

New Camera Technologies. The ubiquity of cameras means that computer-vision algorithms for object detection and tracking are an important topic in Pervasive Computing. A revolution in camera technologies is bringing about a dramatic change in machine-vision capabilities.

Secure Biometrics. Variations in scanned biometrics require approximate matching against stored biometric data for authentication, which seems to rule out encryption of the stored data. Until this problem is solved, biometric applications are dangerously insecure.

Blind Vision & Deaf Audio. What if we could search encrypted images for patterns or objects without ever decrypting them? Similarly, what if we could perform word spotting on encrypted audio? Secure signal processing is a key technology for maintaining privacy while fully exploiting ubiquitous sensors for security and surveillance purposes.

Microfluidic Pumps. A careful power analysis of sensor nodes for several typical environmental-monitoring applications shows that the dominant cost is likely to come from the pumping of reagents and test samples. Reducing the power consumption of small fluid pumps is therefore a key problem in environmental sensor networks.
rf Analog Logic for Low-Power Software Radio. In a typical radio receiver much of the power budget is spent on A-to-D conversion before the DSP even begins its processing. By incorporating data models and signal-processing algorithms into the A-to-D conversion, significant reductions in power consumption may be possible.
Intelligent User Interfaces for Consumer Electronics. While huge advances have been made in the communication and networking aspects of the digital home, almost no progress has been made in the usability of consumer electronics. The next generation of device UIs will be made more effective by giving them the ability to answer the ‘W’ questions: who, what, when, where, why.

Digital Typography for Embedded and Mobile Devices. Hinted outline fonts are less than ideal for the microcontrollers and small memories typically found in embedded and mobile devices. A new shape representation for type has several advantages over the standard representation for ubiquitous computing.

"Proactive Computing"

David L. Tennenhouse
Chief Executive Officer A9.com
www.a9.com

David Tennenhouse is Chief Executive Officer A9.com and formerly vice president of the Corporate Technology Group and director of Research for Intel Corporation.
In addition to building Intel Research, he developed Intel's proactive computing vision, drove several Intel Capital Investments and laid the technical ground work for its new Digital Health Group.

Dr. Tennenhouse is a member of the ACM and a Fellow of the IEEE. Dr. Tennenhouse is currently a Director of the Computing Research Association (CRA), a member of the Advisory Committee of the National Science Foundation (NSF) Computer and Information Sciences and Engineering (CISE) Directorate and a member of the Defense Science Board Task Force on High Performance Microchip Supply. He is also a member of the Dean.s Advisory Board of Carnegie Mellon University.s School of Computer Science and on the Industrial Advisory Board of the Electrical Engineering and Computer Sciences Department at UC Berkeley.

Dr. Tennenhouse previously served on the Visiting Committee on Advanced Technology of the National Institute of Standards and Technology, was a member of the National Science and Technology Council's Sub-committee on Computing Information and Communications R & D, and chaired the Technology & Policy Working Group of the President's Information Infrastructure Task Force. In addition to his journal and conference publications, Dr. Tennenhouse has chaired various workshops and studies concerned with Information Infrastructure, ATM/Gigabit networking, and Advanced Digital Television (HDTV).

Talk Abstract:

This talk will provide a "progress report" on Proactive Computing. For almost 40 years, Computer Science research has been dominated by J.C.R. Licklider's powerful vision of Interactive Computing. Although this "Human Centered" line of research has been tremendously productive, the interactive model will not scale as networked computers begin to outnumber people a hundred or thousand-fold. Simply put, we will not have the human bandwidth to interact with all of those computers. Going forward, we must make better use of Declarative Interfaces, that allow us to simultaneously task large numbers of computers, and Proactive Computing environments, in which networked computers anticipate our needs and, sometimes, take actions on our behalf.