[Published originally in the November 2007 edition of Computing Research News, Vol. 19/No. 5]
Cyber-Enabled Discovery and Innovation
The National Science Foundation (NSF) recently announced its newest foundation-wide, multi-disciplinary initiative, “Cyber-Enabled Discovery and Innovation (CDI),” released as a solicitation http://www.nsf.gov/pubs/2007/nsf07603/nsf07603.htm. In a nutshell, CDI is computational thinking for science and engineering. Computational thinking refers to what the CISE community does in research and education on a daily basis: creating and creatively using computational concepts, methods, models, algorithms, and tools.
CDI seeks revolutionary research in all frontiers of science and engineering enabled by computational thinking. CDI accelerates the huge paradigm shift in the way in which science and engineering will be conducted in the future. It offers a means for tomorrow’s scientists to be trained in and skilled with using computational methods and tools.
CDI projects are to advance more than one field of science or engineering, where computational thinking—that is, “computing” broadly interpreted—counts as one. For years, our community has been reaching beyond computer and information science and engineering, forming intellectual partnerships and establishing the role of computational thinking in other fields of science and engineering. The asymmetric role of computational thinking in the success of CDI gives CISE an instrumental role within the Foundation; indeed, CISE serves as the lead directorate in NSF for CDI. We now have an exceptional opportunity to contribute in ways that are truly transformative for all of science and engineering research and education.
We are drowning in data! The first theme, From Data to Knowledge, focuses on enhancing human cognition and generating new knowledge from a wealth of digital data. CDI seeks to address the fast-growing problem of deriving new knowledge from data sets that are incompatible, heterogeneous, very large, or rapidly flowing. From the use of new computational models to algorithms to tools, scientists and engineers can confirm the expected and reveal the unexpected.
Our systems are growing in complexity. The second theme, Understanding Complexity in Natural, Built, and Social Systems, focuses on deriving fundamental insights on systems comprising multiple interacting elements. Complex systems, from the Internet to atmospheric phenomena, encase human life. This theme promotes the exploration and modeling of natural interactions, connections, complex relations, and interdependencies at all spatio-temporal scales in order to understand, mimic, synthesize, and exploit complex systems.
We work in cyberspace. The third theme, Building Virtual Organizations, focuses on enhancing discovery and innovation by bringing people and resources together across institutional, geographical, temporal and cultural boundaries. Virtual Organizations (VOs) are a vehicle for not only producing transformative outcomes, but also transforming the means of obtaining them. Cyber-based platforms that link together cyber-tools, data sets, and new intellectual partnerships can profoundly change the landscape of research and education in all science and engineering disciplines, including the science of teaching and learning.
CDI welcomes projects that are contained within or cut across these three themes.
Understanding and predicting meso-scale weather, including extremely destructive phenomena such as tornadoes and flash floods, requires high volumes of data that must be assimilated into high-resolution models. The needed input data are diverse, including not only atmospheric parameters such as temperature, pressure and wind speed, but data on land topography, ground cover, locations of bodies of water, and many other variables that can potentially affect local weather. We need networks of smart observational instruments that can adapt their operation modes to changing conditions and alert automated modeling systems to significant events. We need new algorithms to process the incoming data streams in real time and to command the observational instruments in real time. Feeding this data into computational models will produce even more massive amounts of output data to process. We need new computational techniques and tools for visualizing the results so humans can focus their attention on the behaviors of interest, and not be overwhelmed with the inherent complexity of the systems under study. Such research has the potential to transform our understanding of meso-scale weather and enable the accurate prediction of destructive phenomena, all the while requiring new advances in computational thinking.
Predicting weather, of course, is only one scientific grand challenge. Similar examples can be found in astronomy, from discovering brown dwarfs to new galaxies; in the life sciences, from understanding protein structure to plant genomics; in the geosciences, from modeling the inner core to the earth’s surface to the sun; and so on. Computational thinking is key to helping scientists and engineers realize their research dreams.
Even our own field of computing yields its own needs to extract knowledge from data and to understand complexity. Discerning anomalous from normal behavior within a stream of events is a challenge in monitoring a network. In January 2003, the Slammer worm caused a denial-of-service attack infecting 75,000 victims in 10 minutes. In August 2007, a faulty network interface card on a desktop computer caused the computers for the United States Customs and Border Protection Agency to be down for nine hours, keeping more than 17,000 travelers flying into Los Angeles International Airport stuck on planes for hours. A few days later a software flaw shut down the global Skype network for two days, affecting 220 million users. As we become intertwined with “the network” of all kinds of sensing, computing, communicating, and controlling devices, a single failure can have a domino effect that results in large-scale disruptions. These cascading effects are reminiscent of the proverbial butterfly that flaps its wings in the African Sahara and takes the blame for hurricanes on our shores. Advances in network science, drawing from economic theory, multi-scale analysis, and network information theory, can help us model, simulate, and better predict the complex behavior of networks of networks.
CDI views research activities in terms of modes—classifying a project based on its collective intellectual energy and potential outcomes, as opposed to applying simple thresholds to its budget. In its first cycle, CDI will support two types of projects, as defined in the solicitation. Type I projects are efforts led by a small collaborative group of researchers; Type II projects are efforts led by a larger group. In 2009, in addition to Type I and II projects, CDI plans to support Type III projects which are larger, center-scale efforts. The addition of each collaborator to any type of project should have a multiplicative effect on the transformative impact of the project.
All types of projects are encouraged to integrate education and research, where education is interpreted in the broadest sense—from all levels of formal education to the public at large.
For FY 2008, NSF has allocated a minimum of $26M, pending the availability of funds, for CDI. A big team of program officers from every research directorate and programmatic office of the NSF formulated the CDI solicitation and is now ready for the unique challenges of the multidisciplinary review process.
Letters of Intent (LoI) from project team leaders are required and allowed until November 30, 2007. The accuracy of information in the LoI is extremely important in that it will be the basis for NSF’s preparation for the review of the corresponding proposal. Proposals will be reviewed in a two-tier process: preliminary proposals are due by January 8, 2008. Based on the expert panel’s evaluation, successful project teams will be invited to submit full proposals by April 29, 2008. The first cycle of CDI will conclude with awards made by the end of July 2008.
To conclude, we encourage researchers and educators in the computer and information science and engineering community to work with other scientists and engineers, and by putting your collective expertise together, advance computational thinking in the discovery of new science. Send us your creative, bold, and ambitious ideas!
Copyright © 2007 Computing Research Association. All Rights Reserved. Questions? E-mail: email@example.com.