Last Saturday, Erwin Gianchandani, formerly the director of the Computing Community Consortium and now the deputy director of the National Science Foundation’s Division of Computer and Network Systems, organized and moderated a symposium at the American Association for the Advancement of Science’s (AAAS) 2013 Annual Meeting in Boston. The 90-minute session – titled How Fundamental Computing Research Touches Everyday Lives — sought to describe how fundamental research in computing over the last several decades has transformed our world, and promises to facilitate enormous opportunities for still more game-changing breakthroughs in the years ahead. Nearly 85 people packed into a room at the Hynes Convention Center to hear the session’s featured speakers, Jeannette Wing (Microsoft Research), Kevin Knight (USC Information Sciences Institute), and Andrew Lo (MIT Sloan School of Management).
Erwin kicked things off with an overview linking the symposium to the theme of this year’s AAAS Annual Meeting — “The Beauty and Benefits of Science,” and, in particular, the “unreasonable effectiveness” of the scientific enterprise in creating economic growth and solving societal problems — using the “tire-tracks” diagram (right) produced by the National Research Council’s (NRC) Computer Science and Telecommunications Board (CSTB) last summer. The diagram illustrates how fundamental research takes considerable time to pay off, but often pays off in unanticipated ways. Moreover, every one of the multi-billion-dollar sectors of the economy represented in the diagram has a clear connection not just to fundamental research, but also to the complex interplay that exists between academia and industry. Indeed, it is clear that advances in networking and information technology R&D over the last several decades are the result of an innovation ecosystem spanning academia, industry, and government, etc. One story that exemplifies this ecosystem is that of Apple’s iPad. The iPad’s touchscreen keyboard, which has revolutionized the tablet PC business, stems from government-funded academic research throughout the latter part of the 20th century. This research led a pair of researchers at the University of Delaware to establish a company called FingerWorks in 1998, and the company was purchased by Apple in 2005 — just two years before the predecessor to the iPad, Apple’s iPhone, was unveiled.
Jeannette then described how progress in computing technology over the past 20 years has dramatically changed our lives — in our daily routine, at play, and at work. She highlighted advances in mobile devices, wireless networks, robots, sensors and actuators, software services, data analytics, cloud computing, and social media that are behind the conveniences we take for granted today. And she explained how these technologies have had an impact on all other sectors, from healthcare to energy, transportation, agriculture, education, law, engineering, science, and the humanities. Jeannette closed with a powerful video featuring Microsoft Chief Operating Officer Rick Rashid demonstrating a speech recognition breakthrough via machine translation that converts spoken English words into computer-generated Chinese language (below). As Jeannette explained, the breakthrough is patterned after deep neural networks and significantly reduces errors in spoken as well as written translation.
The video was a great segue to Kevin’s talk, which was a “deep dive” into the area of human language technology. Kevin explained that understanding and generating human language is hard for computers because every word has many meanings. The computer must select the right one based on context, but there are an infinite number of contexts. Even then, selecting the right word meanings is not enough — we must then assemble word meanings correctly to understand whole sentences and texts. Kevin went on to describe that, while 15 years ago individual users rarely encountered human language technology, today anyone with web access can use Google Translate (which has its roots in Federally-funded research on statistical machine translation at IBM’s T. J. Watson Research Center), online news aggregators (which have their roots in Federally-funded research at Columbia University), and Apple’s Siri (which has its roots in Federally-funded research at SRI International in Menlo Park, CA). Kevin noted that we have achieved these successes because we stopped worrying about the very hardest linguistic constructions, and instead started worrying about handling the bulk of what computers see and hear. We embraced uncertainty by building probabilistic models of language, he said, and we trained those models on vast data collections. Yet there is much more to achieve: the challenging task of extracting meaning from speech and language will enable many new applications in the future.
Andrew wrapped the symposium with another “deep dive,” this one on the recent financial crisis, which has underscored the complexity of the financial system and the need for a fundamentally different approach to the supervision and regulation of financial institutions. Andrew noted that technological advances in electronic market-making, algorithmic trading, telecommunications, and computing power have created a more efficient financial industry — but one that is highly nonlinear, densely interconnected, and tightly coupled. Consequently, shocks to one node of the financial network can now propagate quickly and unpredictably to other parts of the network. He further described one of the biggest challenges to regulators, the lack of transparency with respect to financial holdings and risk exposures of institutions in the “shadow banking system” comprising financial institutions such as hedge funds, money market funds, and insurance companies. However, forcing these institutions to provide greater disclosure may have the unintended consequence of reducing their incentives to innovate. Andrew illustrated how recent advances in the application of cryptography to create different levels of privacy may provide an ideal compromise between the need for risk transparency and the protection of proprietary information. Indeed, privacy-preserving systemic risk measurement involving big data applications in the consumer finance and banking sectors offers much potential for the future of the financial system.
A key aspect of the symposium was the increasingly interdisciplinary nature of computing. Whether we’re talking about innovating for the smart grid or next-generation transportation networks, or translating natural language, or regulating financial and economic systems, new horizons in computing feature a dual focus on advancing the field of computing all the while simultaneously driving forward other areas of science, engineering, and society. Importantly, as was noted in this space last summer:
This isn’t to suggest that “architecture,” “operating systems,” “compilers,” and “programming languages” are no longer important. Far from it. Rather, as we drive forward advances in interdisciplinary areas like “energy,” “transportation,” “education,” and “health,” so, too, do we drive forward advances at the core of our field.
Ultimately, further advances in computing are essential to tackling the vast majority of our national and global challenges.
The AAAS symposium is an outgrowth of a daylong symposium held in Washington, DC, last February: The Impact of NITRD: Two Decades of Game-Changing Breakthroughs in Networking and Information Technology — Expanding Possibilities Ahead. That event, organized by the CCC and attended by over 150 Federal officials, Congressional staffers, academic researchers, and industry leaders, marked 20 years of the Federal Government’s Networking and Information Technology Research and Development (NITRD) Program, the oldest and largest of the small number of formal Federal programs that engage multiple agencies. In particular, the program has facilitated cooperation and coordination across a broad landscape, providing a framework and mechanisms enabling over 15 Federal agencies to tackle the inherently multidisciplinary, multitechnology, and multisector challenges of today’s networking and information technology R&D horizons. The Program’s success in collaboration has come to be viewed as a model Federal R&D effort that leverages agencies’ strengths and avoids duplication.
Many thanks to our colleagues for participating in Saturday’s session! You can view their slide shows (PDFs) here: Jeannette, Kevin, and Andrew.
