Skip to main content
more options

Overview and Project Highlights

[view as Word Document]

Computational Sustainability: Computational Methods for a Sustainable Environment, Economy, and Society

Institute for Computational Sustainability (ICS), funded under the NSF Expeditions in Computing program, is forging a highly interdisciplinary effort to create and nurture the new field of Computational Sustainability, with the grand vision that computing and information science can – and should – play a key role in increasing the efficiency and effectiveness of the way we manage and allocate our natural resources.  The multi-disciplinary, multi-institutional ICS research team is based at Cornell University and includes leading computer and environmental scientists at Oregon State University, Bowdoin College, Howard University, and The Conservation Fund (TCF).  The ultimate goal of this Expedition is to help alleviate some of the key environmental and sustainability challenges facing our planet today. Concrete examples of potential payoffs are an increased efficiency in the use of our natural resources, more effective computational models for maintaining (or increasing) biodiversity, and large-scale computational equilibrium models for studying trade-offs in the area of renewable energy.  This Expedition has been organized to integrate research and education. The research being conducted by ICS continues to advance discovery and understanding while promoting teaching (through new courses and seminars), training (through workshops and outreach), and learning (through real-world projects and problem cataloging). Given the immediate societal relevance and the environmental component, the area of computational sustainability has found a unique appeal to women and other groups that are traditionally underrepresented in computer science, thus helping increase their participation in the field of computer science.sustainability-triangle

ICS aims to provide solutions for balancing environmental, economic, and societal needs for a sustainable future by bringing computational thinking to sustainability research.  Although a survey of current sustainability literature demonstrates that key sustainability issues translate into decision and optimization problems that fall into the realm of computing and information science, generally they have not been studied by computer scientists. Computational sustainability encompasses problems in disciplines as diverse as ecology, natural resources, atmospheric science, materials science, renewable energy, and biological and environmental engineering. The ICS provides an ideal setting for attacking such problems by bringing together experts from a number of different disciplines.


The research pursued by ICS members is necessarily highly interdisciplinary.  The unique scale, complexity, and impact of such problems have pushed the boundaries of computer science itself, requiring integration of a wide variety of techniques from various areas within computer science and applied mathematics, such as constraint reasoning, optimization, machine learning, data mining, and dynamical systems.  Furthermore, it requires the design and development of models that enable computationally feasible approaches for analyzing systems with highly interconnected components or agents and often with conflicting interests, a necessity for studying and analyzing complex systems that are pervasive in sustainability problems. 

 

ICS members have undertaken a large number of activities to establish computational sustainability as a research area and to demonstrate the effectiveness of computational methods in sustainability research. These activities have thus far resulted in the development of methods, techniques, and models to:

  • Design economically viable conservation corridors for grizzly bears and other species;
  • Develop state-of-the-art methods for predicting bird species occurrence across broad spatial and temporal scales, specifically by taking sampling bias into account;
  • Create new computational analysis methods to help the discovery of new materials, for example for designing more efficient fuel cell technology;
  • Provide new insights into the relationship between fish populations, total allowable catch, and harvest rates;
  • Examine the impacts of increased U.S. gasoline taxes;
  • Evaluate the economic costs, land use adjustments, and greenhouse gas emissions resulting from various biofuels policies;
  • Identify the most promising sites to focus on in order to preserve red-cockaded woodpecker (RCW), a Federally Endangered Species;
  • Understand the impact of climate change in terms of aerosol interactions, desert dust, and paleodata based model estimates;
  • Determine the optimal threshold to commence control strategies for invasive species such as gypsy moths, and to study the effect of invasion in ecosystems involving tamarisk, sudden oak death, and cheatgrass;
  • Improve understanding of feedbacks in the overall system and factors affecting forest fires; and
  • Detect heart rate variability in order to measure stress and ultimately its effect on human behavior.
  • Understand the decision-making processes of pastoral tribes in East Africa.
  • Improve the estimation of poverty levels in Africa and Latin America via novel ensemble machine learning techniques

Research currently underway seeks to use computational methods to:

  • Determine the best wintering and breeding sites to protect to assist the conservation of migratory birds;
  • Studying disease prevention strategies in dairy cows;
  • Minimize the impact of man-made stream barriers on the upstream or downstream movement of migratory fish;
  • Examine the effect of reduced ship strikes on North Atlantic Right Whales and examine the cost of real-time monitoring of ship lanes into major ports and the willingness-to-pay of citizens;
  • Assess potential process configurations for production of biofuels from cellulosic ethanol;
  • Assess the role of carbon offsets in cap-and-trade programs;
  • Develop more robust methods for poverty mitigation through the use of advanced machine learning techniques;
  • Collect and consolidate data from large sensor networks; and
  • Model spatiotemporal herd allocation choice problems of pastoral tribes.

In addition to these application-focused projects, ICS members have made fundamental contributions to computer science and related fields, such as:

  • Deriving new insights into dynamical systems, specifically in the areas of parameter estimation, quantifying uncertainty, and synchronization;
  • Developing effective methods for studying information distribution and fusion in sensor networks;
  • Integrating constraint reasoning / optimization with machine learning methods to develop a scalable hybrid approach for structured unsupervised learning problems;
  • Developing a general methodology for integrating flexible machine learning models into statistical latent variable models;
  • Understanding problem structure in the context of optimization and learning;
  • Develop new classification and feature construction techniques for short time-series data
  • Designing new methods to reason about multi-agent systems in dynamic environments; and
  • Developing large scale computing paradigms for massively large data sets and simulation-intensive studies.

The ICS team has also used Internet websites, conferences, lectures, workshops, publications, and educational video games to promote awareness and understanding of computational sustainability among scientists, policy makers, students, and the general public.


Examples of Research Projects

  • Wildlife Corridor Design for Grizzly Bears and Other Species:  
    Land development often results in a reduction and fragmentation of natural habitat, which makes wildlife populations more vulnerable to local extinction.  One method for alleviating the negative impact of land fragmentation is the creation of conservation corridors, which are continuous areas of protected land that link zones of biological significance.  Researchers have developed a hybrid technique for designing wildlife corridors that yields finer-grained corridor solutions of much higher utility than previous approaches, and often at a fraction of the cost. This work has been driven by and evaluated on the grizzly bear data in the Northern Rockies, with the goal of connecting the Yellowstone, Salmon-Selway, and Northern Continental Divide ecosystems spanning the states of Idaho, Wyoming, and Montana.

    Greater pressures on natural ecological systems caused by factors such as climate change, which will force species to relocate in order to survive, accelerate the complexity of problems such as corridor design and site selection. Artificial intelligence systems such as those developed by ICS researchers, in the hands of wildlife professionals and policy makers, will be an important tool in the face of this complexity for preserving other species and therefore are potentially transformative for our environmental and societal future.
  • Predicting Bird Species Occurrence Across Broad Spatial and Temporal Scales:
    ICS members affiliated with the Cornell Lab of Ornithology and Computer Science department have improved processes for predicting species occurrence across broad spatial and temporal scales. As part of their work related to the Computational Sustainability project, the team released the eBird Reference Dataset, a data warehouse that contains all checklists of observations submitted from the lower 48 states along with detailed location-based covariate information. The team developed a new synthetic analysis workflow to explore and identify interesting patterns that were not previously apparent from species occurrence data. This new paradigm, called Data-Intensive Science, has been the subject of a recently published article in BioScience. The team has also developed nonparametric modeling techniques and model-evaluation techniques that can account for spatial and temporal structuring in species' responses to their environment within a highly effective modeling framework. This analytical method is suitable for species distribution modeling in the face of spatiotemporally varying distributions and habitat associations. The team calls this approach Spatio-Temporal Exploratory Model (STEM), which can be used for intra-annual migration dynamics as well as across year trends.

    The team has demonstrated the utility of STEM analysis for producing models of distributions of migratory and non-migratory species with data from the eBird reference dataset. Recent awards on NSF’s TeraGrid super-computing systems have enabled the team to produce large-scale analysis across hundreds of species and in increasing spatiotemporal resolutions. The results are now being used to inform the government and policy-makers via the State of the Birds report, which studies the distributions of birds across public lands.

  • Site Selection and Species Distribution Models for the Red-Cockaded Woodpecker (RCW):
    Cornell ICS members are working very closely with The Conservation Fund (TCF) team with the goal of locating sites for facilitating the movement of the red-cockaded woodpecker (RCW), a Federally Endangered Species that occurs in at least 10 states. The team developed a draft patch based diffusion model based on cascade models for spread of influence in social networks to describe spatial patterns in RCW populations, and posed this as a stochastic network design problem. The team has projected population/colonization figures that can be used in a more refined optimization model that can incorporate a wide range of constraints. The team has also developed a new integrated mixed integer programming (MIP) optimization model that uses a number of randomly sampled scenarios to combine the stochastic aspects of their patch based diffusion model with land acquisition constraints.

  • Migratory Bird Conservation: Optimal Protection of Wintering and Breeding Sites: The ICS team has made progress towards developing a combinatorial optimization model to determine the best wintering and breeding sites to protect subject to a conservation budget. Conservation of migratory birds requires that habitat be preserved at both wintering and breeding sites, within a limited budget.

  • The Impact of Periodic and Constant Harvesting Policies on 'TAC'-Regulated Fisheries Systems:
    The ICS team has developed a general mathematical model framework for studying the relationship between exploited fish population, total allowable catch (TAC), and harvest rates. Their results provide a strong case against overfishing while highlighting the dangers of global warming. ICS team members also studied periodic harvesting policies, a more general subset of all possible management policies that encompasses constant escapement as a particular case. Their analytical study enables them to computationally find candidate optimal periodic policies for each given period. As a case study on Pacific Halibut data, they have obtained a promising correspondence between theoretical and simulative results.

  • Fish Passage Barrier Removal for Migratory Fish: ICS members working on the problem of minimizing the impact of man-made stream barriers on the upstream or downstream movement of  migratory fish have made progress towards developing a computational approach that is both stronger and yet more flexible than previous approaches. As part of the initial investigation, the team is looking at several culvert network datasets in the State of Washington.

  • Flight Call Detection and Classification:
    ICS members have developed state of the art machine learning algorithms for automatic classification of “Flight Calls.” short sounds produced by birds during migration, by proposing Dynamic Time Warping kernels as an expressive feature construction in conjunction with an explicit multi-class Bayesian probabilistic classifier. The reported classification accuracies, when classifying a flight call to belong to a bird species, are out-performing competing systems and human experts’ levels, and in conjunction with the Cornell lab of Ornithology, the research is now translated and integrated with existing systems for detection and segmentation. ICS research is now looking at ways of fusing additional “meta-data” information from recording stations such as time, location and additional GIS layers. Acoustic detection will provide an additional source of information for bird migration and in conjunction with citizen-science data such as eBird and radar detection it has the potential to explain and monitor large-scale ecological processes.

  • Material Discovery:
    ICS members working with researchers in materials science and physics have found new computational ways of analyzing x-ray diffraction patters in order to help identify new materials with important properties, such as for designing new fuel cell technology.  Using combinatorial methods in materials research has proven to be a powerful approach to the optimization and discovery of inorganic materials. Typically, a library of samples containing a broad range of compositions is evaluated and characterized. The importance of structure-property relationships in such studies has prompted recent developments in high-throughput crystallography. However, it has been a challenge to design a robust algorithm for determining the crystalline phase diagram from a library of x-ray diffraction data. By integrating the actual physics principles underlying the process of experimentation and measurement, ICS members are working on developing innovative ways to enable high throughput investigation of structure-property relationships in complex material systems.  By combining the complementary strengths of constraint reasoning systems on one hand, which are effective at enforcing detailed constraints and inter-relationships, with machine learning methods on the other hand, which are very useful as a data centric approach with a global view, the ICS team has developed a new hybrid methodology that is more scalable and robust than previous approaches.

  • Multi-scale Models for Process Configuration for Biofuel Production, and Equilibrium Policies for Biofuels:
    The ICS team is working on developing multi-scale models for assessment of potential process configurations for production of biofuels from cellulosic ethanol. Their work initially seeks to assess the environmental impacts of a lignocellulosic ethanol fabrication process, using life cycle assessment tools.  The ICS team has also developed a model to evaluate the economic costs, land use adjustments, and GHG emissions resulting from a variety of biofuels policies. The major research finding is that, in contrast with findings from most of the existing research, the increased biofuels mandates do barely bind by 2015; this is because of the implicit path of increased crude oil prices. As a consequence, and in contrast to a recent Science paper (Searchinger et al. 2008), the project participants find that the U.S. economy has the ability to accommodate the production of 15 billion gallons of ethanol by 2015 without disrupting international crop markets. This implies that the unintended land use effects predicted in other papers simply do not exist, and therefore some of the earlier concerns of increased GHG emission from biofuels may have been substantially overstated.

  • The Effect of Reduced Ship Strike on the North Atlantic Right Whale: ICS members have begun a study of the effect of reduced ship strikes on the population structure and viability of the North Atlantic Right Whale. They have made progress in determining the effect of a marginal reduction in the hazard rate on the structure and viability of the North Atlantic Right Whale. The model is being devised to be used to explore the change in population numbers and structure for different ship-strike hazard rates, and will examine the cost of real-time monitoring of shipping lanes into major ports and the willingness-to-pay by U.S. and Canadian citizens.

  • Distributional and Efficiency Impacts of Increased U.S. Gasoline Taxes:
    ICS members and collaborators have examined the distributional and efficiency impacts of increased U.S. gasoline taxes. They found that each cent-per-gallon increase in the price of gasoline reduces the equilibrium gasoline consumption by about 0.2%. Taking account of revenue recycling, the impact of a 25-cent gasoline tax increase on the average household is about $30 per year (2001 dollars). They noted that distributional impacts depend importantly on how additional revenues from the tax increase are recycled.

  • Pastoral Systems in East Africa: ICS participants have been studying various aspects of pastoral systems in East Africa, in particular the modeling of herd allocation choice problem of pastoral tribes in northern Kenya. They have made progress towards providing a method of modeling the pastoralists' spatio-temporal herd allocation choice problem at a level of depth that was not possible with previous methods. The approach is an Inverse-Reinforcement Learning (IRL) framework in which the goral becomes the reconstruction of the pastoral’s decision utilities from the overall observed behavior. ICS members are also involved in the design of the next generation data-collection protocols for pastoral systems that are planned for this year and employ collar-based tracking devices in order to study the effects of introducing Index-based livestock insurance (IBLI). The richness of the collar-based data will allow a detail modeling of decision making via the proposed IRL framework.

  • Poverty Maps: ICS members and colleagues are working towards better ways of designing “poverty maps” to enhance the efficacy of poverty mitigation in less developed countries. The team is developing novel machine learning techniques that are able to improve prediction of poverty levels within a large geographic area by capturing homogeneity of small areas from GIS extracted layers and using such segmentations in order to define sub-models within an ensemble learning approach. The method, called Multiview Ensemble Learning (MEL), combines supervised and unsupervised learning in order to define multiple orthogonal partitions of the geographical space and hence attempt to discover homogeneity in a data-driven manner. The team is performing large-scale experimental comparisons to standard economics approaches and competing data mining methods, such as “bagging,” in order to demonstrate the superiority of MEL when constructing poverty maps.

  • Bioeconomics Meets Biocomplexity: Controlling the Gypsy Moth Population: Institute members have initiated a study of control strategies for invasive species, such as the gypsy moth, by combining perspectives from bioeconomics and biocomplexity. The team has developed a model to determine the optimal threshold to apply a Bt spray to control the damage caused by an established gypsy moth population. This work introduces damage and control costs into a complex biological model and examines the sensitivity of the optimal threshold with respect to key, bioeconomic, parameters. The team has found that the optimal threshold to commence the application of a Bt spray ranges from a spring count of 300 to 450 egg masses per acre.

  • Aerosol Interactions in 3-D Coupled-Carbon-Climate Models: ICS members are working on projecting future carbon dioxide (CO2) levels in order to help predict future climate change patterns. The ocean biogeochemistry, through iron, changes the uptake of carbon dioxide by ~10 ppm over 230 years when desert dust is perturbed by 50-100%, which spans the uncertainties in the desert dust estimates. However, the land responds by taking up more or less carbon in these cases, partly due to changes in precipitation patterns, and partly due to the change in carbon dioxide in the atmosphere. Anthropogenic aerosols change the regional climate, including precipitation and a cooling, causing a decrease in CO2 in year 2100 of about 10 ppm in the team's experiments.

  • Prognostic Fire in the Community Land Model: The ICS team has developed a new algorithm for prognostic fires in the Community Land Model-Carbon and Nitrogen model (CLM-CN) and has shown comparisons to available observations. This model includes the impact of vegetation, soil moisture, wind, and human interventions in the production of fire. This model is designed to be coupled within the Community Climate System Model (CCSM) to allow for the exploration of feedbacks in the system.
  • Heart Rate Variability as a Measure of Stress/Environment Interactions: ICS members are working on using heart rate variability as a measure of stress. One of the greatest challenges to sustainable development is human behavior, and one of the most powerful drivers of behavior is stress.  Recent literature shows that heart rate variability, if carefully analyzed, can be used as a measure of stress. The team has shown that the heart rate monitors can be used to measure heart rate variability in the parasympathetic range (0.15-0.4 Hz). It has also developed preliminary code for best identifying changes in heart rate variability. The long-term goal is to create a new link between the environment, behavior, and neurobiology of stress.

  • Parameter Estimation and Quantifying Uncertainty: Mixed Mode Oscillations in Dynamical Systems: Many problems concerning sustainability naturally involve dynamical systems that are used as predictive models.  The scope of these models ranges from global climate models to models of natural and agricultural ecosystems to models of the energetic efficiency of buildings, transportation systems, and the power grid. ICS members have been working on developing a number of techniques for analyzing dynamical systems. Working on mixed mode oscillations in dynamical systems, ICS members recently extended an earlier analysis of the dynamics associated with folded nodes. ICS members have also constructed new numerical methods for computing invariant slow manifolds.

  • Information Distribution and Fusion in Sensor Networks: Many sustainability problems and solution methods involve the use of sensor networks that monitor the behavior of the system under consideration and continuously provide data to be analyzed for patterns or anomalies.  Such networks are often deployed on a very large scale, for example, throughout a large forested region.  Collecting and consolidating data from such large networks presents many challenges.  ICS participants have considered the belief relationship problem among entities in autonomic networks, which constitutes a major security challenge. This work demonstrates that Markov Random Field theory used in combination with Message Passing algorithms constitutes a powerful theoretical framework for the development of algorithms for information distribution and fusion.

  • Understanding Problem Structure in the Context of Optimization and Learning: ICS members have been exploring ways of understanding and exploiting hidden structure in combinatorial decision and optimization problems, especially in the context of learning during search. For the problem of wildlife corridor design for grizzly bears, the team has identified a clear threshold behavior as well as a matching easy-hard-easy pattern, not previously observed in problems that combine both satisfaction and optimization aspects.

  • Learning Optimal Subsets with Implicit User Preferences: Project participants have been studying the problem of learning a subset of items from a larger ground set from implicit preference requirement of users. They have developed a general purpose Structural Support Vector Machine (SSVM) approach that optimizes a 'set accuracy' performance measure representing set similarities. Once complete, they plan to apply this approach to model fish species distributions, as one can naturally abstract the streams and rivers that various fish species inhabit in a tree-like structure, where each node represents a city or a crossing of rivers, and each edge represents the local closeness of two crossings.  This work also has applications to the spatiotemporal herd allocation choice problem of pastoral tribes in places such as northern Kenya, where each tribe has implicit preferences affected by land topology, available resources, conflicts and affinity with other tribes, etc.

 

Community Building, Policy Guidance, and Outreach
  • International Conferences on Computational Sustainability (CompSust):
    The ICS organized the 1st International Conference on Computational Sustainability (CompSust09) in June 2009 at Cornell University. As the first conference devoted to computational sustainability, CompSust09 allowed computer scientists to interact with representatives from a number of other disciplines and to better understand how those disciplines are approaching sustainability research questions so that computer scientists can determine where the knowledge and methods of their discipline can best be used to speed the path towards a sustainable existence. As a result of CompSust09, a number of computer scientists and their students have developed an interest in working on sustainability-related research, and in further developing a computational sustainability research community.



    One result of this successful outreach was that conference participants from the Massachusetts Institute of Technology (MIT), took an interest in hosting the Second International Conference on Computational Sustainability (CompSust'10). ICS members teamed with these colleagues at MIT to organize CompSust'10, which was held on the MIT campus from June 28-30, 2010. CompSust’10 increased outreach to graduate students and postdoctoral researchers through the introduction of a Doctoral Consortium as part of the conference program, ensuring the vibrancy and “sustainability” of the computational sustainability research community by reaching young scientists early in their careers.
  • Computational Sustainability and Artificial Intelligence: A Special Track of the 25th Conference on Artificial Intelligence:
    The 25th Conference on Artificial Intelligence (AAAI-11) featured a special track on Computational Sustainability and Artificial Intelligence (CompSustAI). The CompSustAI Special Track Cochairs were ICS Director Carla Gomes of Cornell University, and Brian Williams of the Massachusetts Institute of Technology. The conference, held August 7-11, 2011 in San Francisco, CA, brought together Artificial Intelligence (AI) researchers, practitioners, scientists, and engineers in related disciplines. The CompSustAI special track demonstrated growing recognition within computer science that Artificial Intelligence can play a key role in addressing challenges in computational sustainability. The AAAI-11 Outstanding Paper Award went to a paper from the CompSustAI special track.

    State of the Birds 090319 KenhiresThe presentation of research papers on novel concepts, models, algorithms, and systems, in order to address problems in computational sustainability, will served to highlight the impact AI can have in advancing sustainability. In order to improve outreach to potential participants in CompsustAI,  ICS researchers applied machine learning techniques on large collections of online bibliographies (DBLP) in order to discover the Computational Sustainability community in a data-driven manner. The results identified as key researchers in the field, which were then used by the special track cochairs in order to invite papers from those authors. This network discovery project received the “EMC Big Data Award” in Cornell’s annual Bits On Our Minds (BOOM) competition.

    A second special track on Computational Sustainability and Artificial Intelligence will take place at the 26th Conference on Artificial Intelligence (AAAI-12) to be held July 22-26, 2012 in Toronto, Canada.


  • State of the Birds 090319 KenhiresThe State of the Birds Report:  ICS member Ken Rosenberg served on the national science team for The State of the Birds report released by the U.S. Secretary of the Interior in March 2009. This first ever comprehensive report on bird populations in the United States informed policy makers as well as the general public about the decline of bird species in the U.S. and about types of human intervention that can successfully mitigate declines. This report highlighted that nearly a third of the nation’s 800 bird species are endangered, threatened or in significant decline due to habitat loss, invasive species, and other threats. The report also highlighted examples, including many species of waterfowl, where habitat restoration and conservation have reversed previous declines, offering hope that it is not too late to take action to save declining populations. This first report led to 700+ news stories ($1M “ad value”), 126,000 web page views, a bill to reauthorize Neotropical Migratory Bird Conservation Act at $20M (from $5M), a bill to formally authorize USFWS Joint Venture Program, and $3 Million increase in endangered species funds for Hawaiian birds.

    The success of the 2009 report led the national science team, including Rosenberg, to prepare The State of the Birds 2010 Report On Climate Change, released by the U.S. Secretary of the Interior in March 2010.The report shows that climate changes will have an increasingly disruptive effect on bird species in all habitats, and is cited as an unprecedented effort by the Department of the Interior to address climate change.

    For the 2011 report, State of The Birds 2011: Report on Public Lands and Water, Rosenberg was joined by ICS members Theodoros Damoulas and Daniel Fink. This Cornell team asked to contribute year-round species distribution estimates that required high-performance computing techniques. With the 2011 report�s reliance on high-performance computing techniques, the report also serves as evidence of computational sustainability�s roll in addressing sustainability challenges, including species conservation and management of public lands.

 

  • Educational Video Game Development:
    ICS team members and students at Cornell University are developing educational video games intended to enhance the interest in and understanding of computational sustainability among young people and non-scientist members of the public. The games will take players through an experience to solve energy or environmental challenges representative of problems currently being investigated under this grant. In the process, players are empowered by learning about and then 'earning' computer science tools that can be used to analyze and solve the games' challenges. These games are also being developed with the intention that students may be able to take copies of the game home or teachers may use the games in their classrooms.

 

  • CROCS: International Workshops on Constraint Reasoning and Optimization for Computational Sustainability:
    Beginning in 2009, the Institute for Computational Sustainability inaugurated International Workshops on Constraint Reasoning and Optimization for Computational Sustainability (CROCS). ICS members, in collaboration with scientists from Microsoft Research Cambridge, organized the first workshop, CROCS-09 on September 20, 2009 in Lisbon, Portugal, held in conjunction with the 15th Conference on Principles and Practice of Constraint Programming (CP 2009). Building upon the success of this first workshop, ICS members collaborated with scientists representing institutions n the US, Australia, Canada, Ireland, Norway, Spain, and the UK to organize the second International Workshop on Constraint Reasoning and Optimization for Computational Sustainability. This workshop was held on June 15, 2010 in Bologna, Italy, in conjunction with the 7th International Conference on Integration of Artificial Intelligence (AI) and Operations Research (OR) techniques in Constraint Programming (CPAIOR-10). The third International Workshop on Constraint Reasoning and Optimization for Computational Sustainability was organized in conjunction with the 16th International Conference on Principles and Practice of Constraint Programming (CP-10). This workshop was held in St. Andrews, Scotland, UK, on September 6, 2010.

    These workshops brought together interested researchers in order to facilitate the exchange of ideas, presentation of recent or preliminary results, and discussion of promising directions for the use of computational methods to tackle a variety of challenging sustainability problems. The workshops sparked interest in the computer science and operations research communities, especially in scientists working in the fields of constraint reasoning and optimization.

  • sp-illustration.pngScience Pipes:  ICS members affiliated with the Cornell Lab of Ornithology have develop the Kepler-based Science Pipes web application (SciencePipes.org), designed to provide an environment in which students, educators, citizens, resource managers, and scientists can create and share analyses and visualizations of biodiversity data. It is built to bring the benefits of scientific workflows to non-professionals within a simple web environment. Science Pipes allows analysis results and visualizations to be dynamically incorporated into web sites (e.g. blogs) for dissemination and consumption beyond SciencePipes.org itself. Users, from students and teachers to citizen scientists, are excited by the power that Science Pipes delivers to allowing novel analysis of real datasets.

 

 

TRANSFORMATIVE RESEARCH EFFORT


The activities of the NSF-funded Institute for Computational Sustainability (ICS) have been transformative for computing and information science as well as for its key application area of sustainability, across various dimensions:
  1. The ICS research team is nurturing a new research area, Computational Sustainability, with the overarching goal of studying and providing solutions to computational problems for balancing environmental, economic, and societal needs for a sustainable future. Such problems are unique inscale, impact, complexity, and richness, often involving combinatorial decisions, in a highly dynamic and uncertain environment, offering challenges but also opportunities for the advancement of the state of the art in computing and information science.
  1. ICS has integrated in unique ways various areas within computer science and applied mathematics: constraint reasoning, optimization, machine learning, data mining, and dynamical systems.  Furthermore, ICS members have developed models that enable computationally feasible approaches for analyzing systems with highly interconnected components or agents, a necessity for studying and analyzing complex systems that arise in sustainability problems.

As a concrete example of integration, their work on material discovery has brought together techniques from two traditionally disparate sub-fields of computer science, namely machine learning and constraint reasoning.  By bridging the gap between the two sub-areas, ICS researchers have developed a hybrid methodology that is more robust and more scalable than alternative approaches.  As another example, researchers working on policies for fisheries, specifically on TAC-regulated fisheries, and on management policies for bird conservation, specifically for the Red Cockaded Woodpecker, have brought together concepts from the fields of dynamical systems and biological/ecological models on one hand and from inference and optimization on the other.

  1. ICS has fostered a cross-fertilization of approaches and ideas between several research communities through its highly interdisciplinary research team, bringing together computer scientists, biologists, environmental scientists, biological and environmental engineers, mathematicians, and economists from seven different colleges, in twelve different departments at Cornell, Bowdoin College, the Conservation Fund, Howard University, Oregon State University, and the Pacific Northwest National Laboratory.

A central focus of ICS during its first three years has been the establishment of a vibrant new research community in this new area of computational sustainability.  The research and outreach efforts of ICS members (such as organizing the First and Second International Conferences on Computational Sustainability in 2009 and 2010, respectively) have actively promoted awareness and understanding of computational sustainability, while demonstrating through specific research projects that computer science can indeed play a critical role in studying and providing solutions to problems related to the balancing of environmental, economic, and societal needs for a sustainable future.  As a necessarily interdisciplinary endeavor, computational sustainability research is bringing together scientists and other professionals from a wide range of backgrounds to work together in a collaborative process.  These interactions are forging a new community where sustainability researchers can more effectively leverage knowledge and methods beyond their respective disciplines to more rapidly and effectively develop solutions to sustainability problems.

In addition to the transformative nature inherent in establishing a new research area, the efforts of the ICS possess the potential to influence the transformation of human society itself towards a sustainable model of interaction with the natural environment.  Humanity’s use of Earth’s resources is threatening our planet and the livelihood of future generations.  The ultimate goal of ICS and its research is to help alleviate some of the key environmental and sustainability challenges facing our planet today.  The knowledge and solutions developed through computational sustainability research hold the potential to transform the lives of people and species throughout the world.

From a global view, the NSF-funded Institute for Computational Sustainability (ICS) has been a pioneer in bringing advanced computational thinking into the way we manage and allocate our natural resources, and address challenging problems arising in the arena of sustainability.  The advancements in communication and computation in the last two decades have already transformed traditional computational models and provided exciting opportunities, with the emergence of new paradigms and concepts such as electronic markets, just-in-time manufacturing, combinatorial auctions, and customer data mining.  Unfortunately, the impact of information technology has been highly uneven, mainly benefiting large corporate firms in profitable sectors, with little or no benefit in terms of the environment.  On the other hand, governments, international organizations, scientists, and individual citizens are alarmed by climate change, the dramatic increase in our reliance on natural resources, and the unsustainable nature of our current practices.  A survey of the sustainability literature reveals that several key sustainability issues translate into decision and optimization problems that fall into the realm of computing and information science, even though in general they are not studied by computer scientists. These problems range from wildlife preservation and biodiversity to balancing socio-economic needs and the environment to large-scale deployment and management of renewable energy sources to sustainable community design to discovery of important new materials. A detailed study of these problems by ICS members, along with close interactions with domain experts in these areas, has revealed the pervasiveness of the applicability and need for a whole variety of computer science techniques, ranging from combinatorial reasoning and optimization to machine learning and data mining to advanced human-computer interfaces and visualization techniques, to name a few.

Given the well-recognized need for better management and utilization of Earth’s rapidly depleting resources, the activities of ICS have been guided by the principle that it is imperative and urgent for advanced computer science techniques be applied to address computational problems that arise in the context of sustainability. Such problems are unique in scale, impact, complexity, and richness, often involving combinatorial decisions, in a highly dynamic and uncertain environment, offering challenges but also opportunities for the advancement of the state of the art in computing and information science. The work of ICS has integrated in unique ways various areas within computer science and applied mathematics: constraint reasoning, optimization, learning, and dynamical systems. Furthermore, ICS researchers have developed models that enable computationally feasible approaches for analyzing systems with highly interconnected components or agents.  This work has fostered a cross-fertilization of approaches and ideas between several research communities, bringing together computer scientists, biologists, environmental scientists, biological and environmental engineers, mathematicians, and economists from seven different colleges, in twelve different departments at Cornell, Bowdoin College, the Conservation Fund, Howard University, Oregon State University, and the Pacific Northwest National Laboratory.  The effort of ICS members during the first three years of the Institute has already advanced knowledge and understanding within computing and information science while also addressing pressing societal and environmental needs.

 

BROADER IMPACTS


Underrepresented groups:

Through its activities, the ICS has broadened the participation of underrepresented groups. In addition to its director, Carla Gomes, over a dozen female faculty, staff, and student researchers are working to develop the new field of computational sustainability. ICS team members are also varied in their ethnicity and national origins. The ICS sponsored three female graduate students from Cornell's Department of Computer Science to organize workshops for the Expanding Your Horizons (EYH) event, a one-day conference for 7th-9th grade girls. The EYH events were held April 25, 2009 and April 16, 2011on the Cornell University campus. The workshops provided the graduate students with experience in developing curriculum and educational handouts, and presented positive female scientist role models for younger female students. ICS member Abdul-Aziz Yakubu worked to promote the interest of minority students in computational sustainability through talks he gave at two Historically Black Universities, Xavier University in New Orleans and Grambling State University in Grambling, Louisiana. Also, the ICS sponsored a lecture by Margaret Wertheim, who received international acclaim for the 2007 IFF "Hyperbolic Crochet Coral Reef" project. The underrepresentation of women in science and engineering careers was particularly highlighted in Wertheim’s lecture at Cornell, in which she demonstrated the often pent-up excitement women outside of science careers have for learning about and interacting with the sciences through projects like the Hyperbolic Crochet Coral Reef. The excitement was such that the Tompkins County Public Library and Cornell professor Daina Taimina hosted a community workshop to create hyperbolic crochet coral reefs on May 16, 2009 in Ithaca, NY.

Benefits to Society:

While outreach and educational efforts undertaken by ICS members focus on the computational aspects of sustainability problems, these activities also contribute to the public's understanding of challenges facing our society, economy, and environment, and inspire interest and commitment to seeking effective solutions for these problems. ICS members are working very closely with organizations such as The Conservation Fund, the Nature Conservancy, U.S. Fish and Wildlife Service, etc., in order to ensure that their research activities and findings are directly communicated to experts in charge of wildlife preservation, conservation planning, fisheries management, and other impact areas. Many of the sustainability problems being addressed by ICS members have originated from discussions with field-based experts and are being closely guided by them so that meaningful, realistic solutions to complex sustainability problems can be arrived at. Society will by necessity benefit when better solutions for balancing environmental, economic, and social needs for a sustainable future are developed through this new research endeavor.

Infrastructure:

The ICS team’s contributions to infrastructure for research and education include:

  • Development of a community web portal, http://www.computational-sustainability.org, is being developed by the ICS team in order to help build a world-wide community of researchers, educators, policy makers, and students interested in computational sustainability. The web portal will provide everyone with easy, free access to information related to the latest in computational sustainability, such as benchmarks, open-source software, catalog of challenge problems, bibliography, announcements and news, links to publications, etc.
  • Creation and maintenance of the ICS website, http://computational-sustainability.cis.cornell.edu, publicly provides information related to the Institute and allows access to educational materials connected to the Institute's work. The website informs researchers and educators about events sponsored by the Institute which may be of use to the viewer's research or teaching efforts. Presentation slides and video recordings of past sponsored events, such as the 1st International Conference on Computational Sustainability, are also available for these purposes through the website.
  • Release of the eBird Reference Dataset by ICS members affiliated with the Cornell Lab of Ornithology (Munson et al. 2009).  It is available at the AKN website, and is freely available for all usages and is designed to allow researchers to benchmark their experiments in a standardized fashion.  The Reference Dataset is a data warehouse that contains all checklists submitted from the lower 48 states of the United States along with detailed location-based covariate information.  It contains information regarding the sampling event, environmental covariates that are felt to be generally useful for modeling species occurrence, and a series of spatial pattern statistics.
  • Development of museum exhibit and 'edutainment' video games which are being developed by project participants and collaborators to be displayed at science and technology museums, educational events, and in other such venues. This effort has been highly educational for the more than 30 Masters and Undergraduate students themselves who are working on developing these 5 edutainment games.  The exhibit would inform attendees of the creation of the Computational Sustainability field and motivate them to view computer science as a powerful tool in helping to meet our society's energy and environmental needs. Additionally this exhibit aims to use this message to excite young students about computer science as a way that they too can make a difference in the world. The 'edutainment' video games, under development and active testing at middle schools in the area, would enable students to learn about and then 'earn' computer science tools that can be used to analyze and solve the games' challenges. These games are also being developed with the intention that students may be able to take copies of the game home or teachers may use the games in their classrooms. They will also be made available online in order to expand the reach of this educational tool.

Teaching, Training, and Development:

ICS activities have been in a new area for most of the participants, providing them a very different research and learning experience. Many computer science and applied mathematics students, and even experienced researchers, have for the first time found the opportunity to work very closely with field-based experts from organizations such as The Conservation Fund, U.S. Fish and Wildlife Service, U.S. Department of Agriculture, etc. This is a unique learning experience for ICS members, especially for those with a background in computer science related fields where the university curriculum as well as basic research is fairly abstract and somewhat far from immediate application to pressing environmental and economic issues in the society. ICS members are continuously learning to interact with experts from other fields, and are exploring new ways to apply their computational skills to help tackle complex sustainability problems. Additionally, graduate and undergraduate students have received mentorship in the course of their research from the faculty who oversee their work with various research activities. ICS team members have developed their own knowledge and skills and provided opportunities for others to do so as well through sponsorship of research visits for students and researchers from other institutions, their presentations at seminars, conferences, and workshops, their organization of the 1st Conference on Computational Sustainability, and development of the ICS and computational sustainability websites.

Dissemination:

Results of research conducted by the ICS will be disseminated broadly to enhance scientific and technological understanding. This dissemination draws upon the infrastructure being created through development of the computational sustainability and ICS websites, which will make information and new applications available to anyone with internet access regardless of geographic location. Dissemination is already occurring and will continue to occur through publications, lectures, and papers presented by ICS team members, annual computational sustainability conferences, educational and outreach events, and development of a new journal for the field.

Drop us an email if you'd like to join us in establishing the field of Computational Sustainability.

NSF

Sponsored by the National Science Foundation