All posts by douglashfisher

Director of Outreach, Education, Diversity, and Synthesis (OEDS) of CompSustNet

A Review of Wildbook: Software to Support the Environment

This is a guest post by Carmen Camp. See Carmen’s bio at the bottom of this post

Wildbook is an open source software platform devoted to tracking animals in the wild and decreasing the chances of extinction in different species. The software itself, WildMe, is publicly available on GitHub to encourage people to join the movement in protecting wildlife. Much of the success of the program relies on citizen science and interaction, and the system itself utilizes artificial intelligence to identify animals and other environmental features from photographs. A unique advantage of using photos for identification purposes is that when an individual animal is encountered in the wild, today’s digital cameras are able to capture a large number of high quality photographs before the encounter is over. These photos may then be used to further improve animal identification by using machine learning software that is available through Wildbook.

Wildbook processes photos from research teams, social media, and citizens alike. Wildbook uses deep convolutional neural networks to analyze a photograph, and identify animals, plants, and other objects contained within the scene. The system’s pattern matching algorithms are also capable of identifying unique individuals that are known to research teams. This gives scientists important opportunities to track individuals, as well as populations, and social interactions of the animals.

Early Work with Whale Sharks

The purpose of Wildbook is to be an open-source platform that enables different research teams to perform photo identification with limited manual labor. Since its beginnings, Wildbook has been used by many projects, one of the first of which was a 2005 project to study whale sharks. The photo identification software was based on software to identify constellations (Arzoumanian et al., 2005). Once adapted, the software enabled researchers to identify whale sharks from their size and shape, as well as their spot patterns. Spot coordinates are represented in XML and either match known individuals, or are used to identify previously unrecorded individuals.

A later study in 2008 also used a whale shark database along with the lab’s mark-recapture methods to better understand the survival rates of whale sharks in Western Australia (Holmberg et al., 2008). The 2005 and 2008 studies, and others, have led to a “Wildbook for Whale Sharks” organization that is dedicated primarily to keeping track of whale sharks around the world. The organization’s website has a featured link that allows users to “Report your sightings,” which requests input about the user, the sighting, its date and location, and whatever footage or photos the user captured of the whale sharks.

Improvements in the Wildbook for Whale Sharks software has spread coverage of whale sharks from Western Australian populations, to the Philippines, to the Western Atlantic Ocean. While some studies focus on geographic areas and the distribution of the animals (Araujo et al., 2016), others track specific animals (Norman and Morgan, 2016). Still other projects track social groups and seasonal migrations of different populations, like the 2013 study on sharks in the Gulf of Mexico and the Caribbean Sea (Hueter et al., 2013).

Extending Wildbook to other Species

Not only can an organization for an individual species like whale sharks use Widbook, but citizens and scientists alike can participate in the research by providing photographic evidence of many types of animals. Manta Matcher is another example of programs that have been able to form because of the capabilities offered by Wildbook. Its website has a nearly identical setup to Wildbook for Whale Sharks, and it provides easy access to a page through which users can submit data and photographs they collected. Flukebook is yet another option for photo identification that is, as its name might suggest, an identifying software for the flukes of whales. Each of these programs that are directly related to Wildbook offer resources that in turn are used by many scientific organizations. For example, the Dominica Sperm Whale Project uses Flukebook to keep track of individuals in the island’s surrounding waters. Shane Gero, a scientist from the project, is quoted on WildMe’s website, saying, “PhotoID as a tool for conservation and research finds power in numbers and international, inter-institutional collaboration,” indicating that a globally collaborative platform like Wildbook is exactly what the scientific world needs to get answers and solve problems.

Some programs use Wildbook to find individual known animals, rather than categorizing creatures as one species or another. In 2013, an algorithm was introduced called HotSpotter, which used pattern matching methods to identify unique sections on photographed animals. Thus, HotSpotter became a versatile option for identifying multiple types of animals, as well as individuals within the same species. HotSpotter focuses on identifying key points on animals in the frame of the photograph, and it then uses a nearest-neighbor search to compare the new photo with pre-existing records in the database of individual animals. To train and test the system, it was run on photos from scientists, assistants, ecotourists, and ordinary citizens. The software has been successfully used on many creatures, including two types of zebras, giraffes, leopards, and lionfish (Crall et al., 2013).

Citizen Scientists and Wildbook

Every example of Wildbook usage for external research highlights the importance of citizen interactions that make the research possible. Machine learning schemes and artificial intelligence typically require large amounts of data with which to be trained, and in order to test the algorithms, completely fresh, non-repeated data files must be used. This means that in order for an algorithm to be effective and accurate, it must have a huge source of data files, which in this case are photographs. Thus, citizens’ and the public’s interactions with projects like Wildbook are essential to success. This necessity for external input is evident in many related pieces of research, such as a 2017 case study on Twitter. This study discusses the crucial role that everyday members of the public play in data collection. In this study, the researchers wanted to train a machine learning algorithm to identify emotions in tweets on Twitter. The people on the research team, however, could not manually provide enough examples and training data to have a fully functioning algorithm. It simply would take far too long to be worthwhile. Using citizen scientists, however, the team was able to gather enough data to get their algorithm running accurately in a reasonable amount of time (Sastry et al., 2017).

Not only is advancing technology providing more opportunities for the public to get involved with research, but it is also offering new and more accessible ways of participating in different projects. Take, for example, the Humane Society, which asks citizens to submit information about roadkill they come across online. The National Audubon Society also has a program for volunteers to count and monitor birds for an annual census. Programs such as these, as well as Wildbook spinoffs, allow citizen scientists to submit vital information through the internet or applications. Evolving technology has placed the ability to submit such data directly in our hands through mobile phones and other devices, and the internet offers places to submit data, as well as aids to citizens in finding causes and projects with which they can help.

WildMe’s website divides citizen scientists into four distinct groups that are categorized by the role the citizens play in research. The first group are denoted as “scientists.” These citizens are incredibly engaged in the research, and they focus their efforts on analyzing data and determining its meaning. Second are the “evangelists,” who are devoted to outreach and explaining the research project to the public. They play an important role in motivating more people to join them in the effort, as well as building communities that support the research. The next role is “the technologist,” which further emphasizes the significance of technological advancements. These people make sure that the IT side of the project allows it to be as efficient and interactive as possible. Finally, “the volunteer” educates members of the public so that they are capable of collecting data, monitoring inputs, or analyzing information, as it may relate to the research.

Wildbook in the Global Community

Those in charge of Wildbook also understand the power of globalization and how the world is connected through the internet. With an active Twitter account, the organization is able to advertise itself through new pieces of research that use the software. The Wildbook Facebook page is also active. It provides similar updates to Twitter, as well as chances for visitors to donate to the cause, attend related events, and participate in virtual reality activities to learn more about the animals. For example, The “Great Grevy’s Rally” is currently publicized on the Facebook page, and it invites people to go to Kenya in January to help complete a census on the Grevy’s zebras in the area. It welcomes any and all aspiring citizen scientists to join the charge in driving around a designated area to photograph and document and zebras seen in the area. The data collected will then be put into the Grevy’s zebra Wildbook database. The Facebook link on the event page redirects one to a page describing the Grevy’s zebra mission, and clearly offering tips on how to become a citizen scientist and help the cause. By marketing these events in places that supporters are likely to see them, Wildbook gains both support and renown via the internet.

Wildbook provides an incredible opportunity to globalize scientific and ecological missions like was never before possible. Individuals of any profession from around the world can participate in the global mission to save and preserve the planet on which we live. The software provides an interface between science and the public that is easily accessible, especially for individual species that have their own associated organizations and easily accessible databases and websites. From whale sharks, to giraffes, to leopards, to lionfish, Wildbook has introduced endless options for collaboration. More than simply using user-generated photographs, Wildbook offers people the option to get involved and have a part in science, which is a crucial piece of gathering a force together that can have a positive impact on our changing world and dangerously shifting animal populations.

Carmen Camp will graduate in spring 2018 with a degree in Computer Science and Corporate Strategy. She is passionate about marine science and hopes that her future will include plenty  of opportunities to help protect the ocean. She may be contacted at


Araujo, G., Snow, S., So, C. L., Labaja, J., Murray, R., Colucci, A., and Ponzo, A. (2016) Population structure, residency patterns and movements of whale sharks in Southern Leyte, Philippines: results from dedicated photo-ID and citizen science. Aquatic Conserv: Mar. Freshw. Ecosyst., doi: 10.1002/aqc.2636.

Arzoumanian Z, Holmberg J & Norman B (2005) An astronomical pattern-matching algorithm for computer-aided identification of whale sharks Rhincodon typus . Journal of Applied Ecology 42, 999-1011.

Hueter RE, Tyminski JP, de la Parra R (2013) Horizontal Movements, Migration Patterns, and Population Structure of Whale Sharks in the Gulf of Mexico and Northwestern Caribbean Sea. PLoS ONE 8(8): e71883. doi:10.1371/journal.pone.0071883

Holmberg J, Norman B & Arzoumanian Z (2008) Robust, comparable population metrics through collaborative photo-monitoring of whale sharks Rhincodon typus. Ecological Applications 18(1): 222-223.

J. P. Crall, C. V. Stewart, T. Y. Berger-Wolf, D. I. Rubenstein and S. R. Sundaresan, “HotSpotter — Patterned species instance recognition,” 2013 IEEE Workshop on Applications of Computer Vision (WACV 2013), pp. 230-237.

Norman B. and Morgan D. (2016) The return of “Stumpy” the whale shark: two decades and counting. Front Ecol Environ 2016; 14(8):449–450, doi:10.1002/fee.1418

Sastry, N., et al.: “Bridging big data and qualitative methods in the social sciences: A case study of Twitter responses to high profile deaths by suicide,” Online Social Networks and Media (2017)

The Importance of Citizen Scientists


RegionRadio and Related Works

This is a post by Emily Markert. See her bio at the bottom of this post.

Many of us have experienced car rides, both short and long, filled with small-talk, music played from a radio, or even podcasts. Although these media seem to do well enough at keeping us entertained, they often contribute little to any higher purpose. The vast amount of time people spend in transit has great potential to provide a meaningful experience, yet typically goes unutilized. Thus, we introduce RegionRadio, a spatially-aware storytelling tool aimed at supporting sustainability efforts through place-based education. The theory behind place-based education argues that immersing people in the history of a place can increase the connection they feel to it, and therefore increase the likelihood that they would act to protect it. In this way, RegionRadio exhibits place-based education by presenting stories to a user as they travel along a route; these stories are mandated to be relevant to the user’s current geospatial location, to contain topics related to environmental concepts, and eventually, to be customized to specific users’ preferences and histories. These stories are analyzed and filtered based on their topic, perceived interestingness, and continuity of the ‘playlist’ presented to the user throughout their journey. Use of RegionRadio is intended to be as simple as turning on the radio, but aims to develop ordinary citizens as environmental advocates by heightening appreciation of their surroundings through the telling of stories.

Although this project is unique in its sourcing and presentation of environmentally-focused place-based education,  we take inspiration from other work in spatially-based education, to include SCRABS and DETOUR.

The SCRABS system was designed by a group of researchers from the Cultural Heritage Information Systems national project, and is intended to increase cultural appreciation by presenting users with personalized, context-based information as they explore historic sites. The system has not yet been fully developed, but presents an ambitious design that can be further reviewed in this publication. This context-dependent recommendation of information aimed at increasing cultural appreciation is highly relevant to our own task, but applied to a different domain.

We consider another tool that is fully-developed, and has potential to convey place-based education. Detour is a growing platform offering the creation and taking of a novel type of audio tour. Taken on the user’s own smartphone, Detours can be taken in both indoor and outdoor spaces, and progress as the user moves through space, triggering narrations and other media along the way. While most existing tours published with Detour focus on urban areas, there is great potential in creating Detours for more natural settings. Allowing users to download tours that might lead them on a hike, through a National Park, or on a scenic drive would make place-based education easily accessible for Detour’s thousands of users.

To keep up with RegionRadio, stay tuned to this blog.

Region Radio a collaboration between the CompSustNet  lab at Vanderbilt University and the Space, Learnng, and Mobility lab at Vanderbilt University.  Research and development of Region Radio is supported by NSF Award #1521672 “Collaborative Research: CompSustNet: Expanding the Horizons of Computational Sustainability” and NSF Award #1623690 “EXP: Bridging Learning in Urban Extended Spaces (BLUES) 2.0

Emily Markert is a Computer Science undergraduate at Vanderbilt University. The opinions expressed herein are Emily’s and not necessarily those of Cornell University. You can reach Emily at

Creating Citizen Science Projects

This is a post by Mateus Winelmann. See Mateus’ bio at the bottom.

We often hear about major scientific discoveries in the media, like how scientists at CERN found the Higgs Boson or that a lifesaving drug like Harvoni, which can cure hepatitis C, is going to market. These are amazing breakthroughs, but oftentimes the process of discovery feels like something far removed from us. After all, most of us don’t have the training or resources to work on these kinds of projects. When cutting edge scientific research often requires years of specialized education and billions of dollars in funding (finding the Higgs Boson is estimated to have cost over US$13 billion), it doesn’t seem like something ordinary people can be a part of. That doesn’t have to be the case though! Citizen science consists of research done largely by the general public, typically without any significant cost for participants, and it can be incredibly valuable. To illustrate that, let’s talk about eBird.

eBird was launched in 2002 by the Cornell Lab of Ornithology and the National Audubon Society, where the idea is to allow people to document their bird sightings. Bird watchers already tend to keep records of the birds they see or hear, and eBird allows them to make those observations available to educators and researchers across the world. Making that data available is already bearing fruit, with ornithologists at Cornell publishing a paper last year discussing the migratory strategies of birds, which is discussed in a New York Times article. The paper’s authors were able to document how different species of birds from different parts of the country migrate in different ways thanks to eBird. According to one of the paper’s articles, it would have cost researchers millions of dollars to collect this data through traditional tracking methods, and even then, the data would not have been as detailed.

You can learn more about eBird and how to contribute here. If you are interested in finding other citizen science projects, take a look at NatureNet and SciStarter. If you’re feeling particularly ambitious, you can consider starting your own citizen science project. There are several tools and frameworks out there that can help you create and share a citizen science project, some of which are described here, and a few others I would suggest looking at are CitSci, crowdcrafting, Zooniverse, and iNaturalist. Starting your own project can be an ambitious undertaking, so I’d recommend taking the time to explore what projects are already out there to get an idea about how you might create your own project. While this page is directed at educators, it highlights a few important things to consider if you decide to create your own project. There is no shortage of things to be studied, so it’s just a matter of finding, or starting, a project that interests you.

Mateus Winelmann is a senior undergraduate student at Vanderbilt University. The opinions expressed herein are Mateus’s and not necessarily those of Cornell University. You can reach Mateus at

Livestock Insurance in Africa

This is a post by Emily Markert. See her bio at the bottom of this post.

For pastoralists in East Africa, weather is key.  The threat of drought is recurrent on the African rangelands, and has the potential to kill vast numbers of livestock, throttling herders and their families into poverty.  This uncertainty has been a long-term concern in the region, and has led researchers Andrew Mude, Chris Barrett, and Michael Carter to develop a technology-based insurance program to protect these herders.  This Index-Based Livestock Insurance uses satellite data to monitor weather conditions in pastoral regions, and estimates livestock deaths.  Herders receive payouts based on these predictions.

A team led by Carla Gomes, Director of the Institute for Computational Sustainability, has also developed mobile applications that allow herders to report conditions, introducing an element of citizen science.  The program has been implemented in multiple countries, and this innovative combination of technology and finance has proven to be a success.  The researchers behind this insurance have received numerous awards for their efforts, and countless pastoralists have seen their livelihoods stabilized.  More information on this program can be found in this article from the Cornell Chronicle, in a press release on Dr. Andrew Mude’s receipt of the 2016 Norman Borlaug Award for Field Research and Application, or in this blog post.

Emily Markert is a Computer Science  undergraduate at Vanderbilt University. The opinions expressed herein are Emily’s and not necessarily those of Cornell University. You can reach Emily at

Awards for Computational Sustainability Papers at AAAI-17 and IAAI-17

Computational sustainability has been a special track at AAAI since 2011. The track invites “research papers on novel concepts, models, algorithms, and systems” at the nexus of AI, and environmental and societal sustainability. The 2017 special track cochairs were Bistra Dilkina of Georgia Institute of Technology and Sabine Storandt of Julius-Maximilians-Universität Würzburg.

There were two CompSust awards given at the 2017 AAAI conference. The AAAI CompSust Best Paper Award was given to
Xiaojian Wu, Akshat Kumar, Daniel Sheldon, and Shlomo Zilberstein for “Robust Optimization for Tree-Structured Stochastic Network Design“. The CompSust Best Student Paper was given to Jiaxuan You, Xiaocheng Li, Melvin Low, David Lobell and Stefano Ermon for their paper “Deep Gaussian Process for Crop Yield Prediction Based on Remote Sensing Data.

The AAAI CompSust Best Paper Committee included Alan Mackworth of the University of British Columbia, Zico Kolter of Carnegie Mellon UIniversity, and Amy McGovern of the University of Oklahoma.

The IAAAI conference, co-located with AAAI, also had computational sustainability representation, and CompSustNet researchers received an IAAI-17 Deployed Application Award for
Phase-Mapper: An AI Platform to Accelerate High Throughput Materials Discovery” by  Yexiang Xue, Junwen Bai, Ronan Le Bras, Brendan Rappazzo, Richard Bernstein, Johan Bjorck, Liane Longpre, Santosh K. Suram, Robert B. van Dover, John Gregoire, and Carla P. Gomes.

Finally,  CompSustNet Executive Council member co-authored  the AAAI-17 Outstanding Paper “Label-Free Supervision of Neural Networks with Physics and Domain Knowledge” by Russell Stewart and Stefano Ermon. The paper describes research on using constraints to reduce the need for labeled data when learning to recognize and track objects. While not a computational sustainability paper per se, it a line of research that has clear implications for computational sustainability.

AAAI-17 CompSust co-chair Bistra Dilkina presents Best Paper award to Xiaojian Wu and Akshat Kumar, for their paper with Daniel Sheldon and Shlomo Zilberstein.


A complete list of computational sustainability representation at AAAI and IAAI is broader than the CompSust special track, and relevant papers from AAAI, IAAI, and the AAAI-17 Workshop on AI+OR for the Social Good (appended at the end) are listed here as a convenience. This list includes CompSustNet members and collaborators too, highlighted in blue.

Sunday, February 5
EAAI-17 Blue Sky Ideas in AI Education from the New and Future AI Educator 
Sun 5-5:50, Golden Gate 1-2
  • AI Education through Real-World Problems by Mark Crowley 
Monday, February 6
AIW1: Crowdsourcing Techniques and Methodologies  
Mon 10-11, Continental 9
  • Poster 1475: Species Distribution Modeling of Citizen Science Data as a Classification Problem with Class-Conditional Noise by Rebecca A. Hutchinson, Liqiang He, Sarah C. Emerson
GTEP1: E-Commerce and Auctions 
Mon 10-11, Golden Gate 6
  • Poster 2250: Proper Proxy Scoring Rules by Jens Witkowski, Pavel Atanasov, Lyle H. Ungar, Andreas Krause
IAAI-17: Transportation: AI Applied to Safer and More Efficient Travel 
Mon 11:30-12:30, Golden Gate 5
  • Risk-Aware Planning: Methods and Case Study on Safe Driving Routes  by John Krumm, Eric Horvitz
  • Predicting Fuel Consumption and Flight Delays for Low-Cost Airlines  by Yuji Horiguchi, Yukino Baba, Hisashi Kashima, Masahito Suzuki, Hiroki Kayahara, Jun Maeno
  • Determining Relative Airport Threats from News and Social Media  by Rupinder P. Khandpur, Taoran Ji, Yue Ning, Liang Zhao, Chang-Tien Lu, Erik R. Smith, Christopher Adams, Naren Ramakrishnan
PS1: Planning
Mon 2-3:30, Plaza A
  • Poster 790: Three New Algorithms To Solve N-POMDPs by Yann Dujardin, Tom Dietterich and Iadine Chadès
VIS3: Object Recognition 
Mon 2-3:30, Golden Gate 7-8
  • Poster 2845: Extracting Urban Microclimates from Electricity Bills by Thuy Vu, D. S. Parker
MLA3: Machine Learning Applications   
Mon 2-3:30, Golden Gate 3
  • 1289: Deep Spatio-Temporal Residual Networks for Citywide Crowd Flows Prediction Junbo Zhang, Yu Zheng, Dekang Qi
IAAI-17: Deployed AI Systems 
Mon 2-3:30, Golden Gate 5
  • Phase-Mapper: An AI Platform to Accelerate High Throughput Materials Discovery  by Yexiang Xue, Junwen Bai, Ronan Le Bras, Brendan Rappazzo, Richard Bernstein, Johan Bjorck, Liane Longpre, Santosh K. Suram, Robert B. van Dover, John Gregoire, Carla P. Gomes
AAAI-17 Invited Panel: AI for Social Good
AAAI  Mon 4-5, Continental 4-6
  • Panelists: Eric Horvitz, Peter Mockel, Lynne Parker, and Gideon Mann. Moderated by Milind Tambe.
PS2: Deterministic Planning 
Tue 10-11, Golden Gate 1-2
  • Matrix Factorisation for Scalable Energy Breakdown by Nipun Batra, Hongning Wang, Amarjeet Singh, Kamin Whitehouse
Senior Member Talks 1 (Summary)  
Tue 11:30-12:30, Continental 9
  • 3441: A Selected Summary of AI for Computational Sustainability by Douglas H. Fisher
ML12: Methods 
Tue 11:30-12:30, Plaza A
  • 788: On Human Intellect and Machine Failures: Troubleshooting Integrative Machine Learning Systemsby Besmira Nushi, Ece Kamar, Eric Horvitz, Donald Kossmann
GTEP6: Game Theory
Tue 2-3:30, Golden Gate 6
  • Poster 23: Algorithms for Max-Min Share Fair Allocation of Indivisible Chores by Haris Aziz, Gerhard Rauchecker, Guido Schryen, Toby Walsh
MLA6: Deep Learning / Neural Networks 
Tue 2-3:30, Golden Gate 3
  • Poster 629: Combining Satellite Imagery and Open Data to Map Road Safety by Alameen Najjar, Shuníchi Kaneko, Yoshikazu Miyanaga
  • Poster 573: Regularization in Hierarchical Time Series Forecasting with Application to Electricity Smart Meter Data by Souhaib Ben Taieb, Jiafan Yu, Mateus Neves Barreto, Ram Rajagopal
ML15: Reinforcement Learning
Tue 2-3:30, Plaza A
  • Oral 2097: Maximizing the Probability of Arriving on Time: A Practical Q-Learning Method by Zhiguang Cao, Hongliang Guo, Jie Zhang, Frans Oliehoek and Ulrich Fastenrath
VIS6: Videos
Tue 2-3:30, Golden Gate 7-8
  • Poster 1854: Counting-Based Reliability Estimation for Power-Transmission Grids by Leonardo Duenas-Osorio, Kuldeep S. Meel, Roger Paredes, Moshe Y. Vardi
IAAI:17: Smart Environments: Using AI Systems to Improve Day-to- Day Life 
Tue 2-3:30, Golden Gate 5
  • Crowdsensing Air Quality with Camera-Enabled Mobile Devices by Zhengxiang Pan, Han Yu, Chunyan Miao, Cyril Leung
  • Real-Time Indoor Localization in Smart Homes Using Semi-Supervised Learning by Negar Ghourchian, Michel Allegue-Martinez, Doina Precup
  • ParkUs: A Novel Vehicle Parking Detection System by Pietro Carnelli, Joy Yeh, Mahesh Sooriyabandara, Aftab Khan
MLA8: Applications of Supervised Learning  
Tue 4-5, Golden Gate 3 
  • 1751: Predicting Demographics of High-Resolution Geographies with Geotagged Tweets  by Omar Montasser, Daniel Kifer
ML17: Classification and Clustering   
Tue 4-5, Plaza A
  • 353: POI2Vec: Geographical Latent Representation for Predicting Future Visitors by Shanshan Feng, Gao Cong, Bo An, Yeow Meng Chee
IAAI-17: Decision Support: AI for Better Decision Making
Wed 10-11, Golden Gate 5
  • Cracks Under Pressure? Burst Prediction in Water Networks Using Dynamic Metrics by Gollakota Kaushik, Abinaya Manimaran, Arunchandar Vasan, Venkatesh Sarangan, Anand Sivasubramaniam
  • Optimal Sequential Drilling for Hydrocarbon Field Development Planning by Ruben Rodriguez Torrado, Jesus Rios, Gerald Tesauro
SCS1: Constraint Satisfaction  
Wed 11:30-12:30, Golden Gate 7-8
  • Poster 1556: General Bounds on Satisfiability Thresholds for Random CSPs via Fourier Analysis  by Colin Wei, Stefano Ermon
RU2: Sequential Decision Making 
Wed 11:30-12:30, Golden Gate 4
  • Poster 2970: Hindsight Optimization for Hybrid State and Action MDPs by Aswin Raghavan, Scott Sanner, Roni Khardon, Prasad Tadepalli, Alan Fern
STCOMPS1: Dynamic and Spatiotemporal Systems
Wed 2-3:30, Golden Gate 4
  • Oral 2042: Fast-Tracking Stationary MOMDPs for Adaptive Management Problems by Martin Péron, Kai Helge Becker, Peter Bartlett, Iadine Chadès
  • Oral 823: Robust Optimization for Tree-Structured Stochastic Network Design by Xiaojian Wu, Akshat Kumar, Daniel Sheldon, Shlomo Zilberstein
  • Oral 3171: Dynamic Optimization of Landscape Connectivity Embedding Spatial-Capture-Recapture Information by Yexiang Xue, Xiaojian Wu, Dana Morin, Bistra Dilkina, Angela Fuller, J. Andrew Royle, Carla P. Gomes
  • Oral 2303: Spatial Projection of Multiple Climate Variables Using Hierarchical Multitask Learning by André R. Gonçalves, Arindam Banerjee, Fernando J. Von Zuben
  • Oral 2412: Deep Gaussian Process for Crop Yield Prediction Based on Remote Sensing Data by Jiaxuan You, Xiaocheng Li, Melvin Low, David Lobell, Stefano Ermon
MLA11: Machine Learning Applications  
Thu 11:30-12:45, Golden Gate 6
  • Oral 1922: Fine-Grained Car Detection for Visual Census Estimation by Timnit Gebru, Jonathan Krause, Yilun Wang, Duyun Chen, Jia Deng, Li Fei-Fei
ML25: Recommender Systems
Thu 11:30-12:45, Continental 1-3
  • Oral 1502: Polynomial Optimization Methods for Matrix Factorization by Po-Wei Wang, Chun-Liang Li, J. Zico Kolter

Workshop on AI+OR for Social Good

14:00-14:35: Invited Talk: Daniel Sheldon, University of Massachusetts Amherst and Mount Holyoke College, “AI for Ecology and Conservation”

14:35-14:55: Sean Mcgregor, Rachel Houtman, Claire Montgomery, Ronald Metoyer and Thomas Dietterich, “Factoring Exogenous State for Model-Free Monte Carlo”

14:55-15:15: Yiqun Xie, Kwangsoo Yang, Shashi Shekhar, Brent Dalzell and David Mulla, “Spatially Constrained Geodesign Optimization (GOP) for Improving Agricultural Watershed Sustainability”

15:15-15:35: Rui Zhang, Jefferson Huang and Tarun Kumar, “Preventive Leak Detection for High Pressure Gas Transmission Networks”

Promoting Computational Sustainability to the Public

This is a post by Selina Chen. See her bio at the bottom of the post.

When I attended the 4th International Conference on Computational Sustainability at Cornell University, I was surprised at the large number of projects that came with some involvement in incentivizing users — whether by making the technology more accessible for the average user or coming up with ways to get people interested in using the product.  Before the conference, I had thought research consisted of the studying and making of things, and the issue of actually “selling” the product to public would be one which would be foisted off to the companies who decided to use the research.  Doubtless, my view was clouded by the fact that that I’d been raised by two biochemists, whose jobs consisted of developing and testing new drugs, which would then be marketed and sold by another company or department in that company.  Yet, as I learned at the conference, this cycle of create-and-pass-off was not so in the CompSust research community. Oftentimes, it became the job of researchers to figure out how to properly integrate sustainability into public life by paying attention to behavioral science, as well as technology.

In her blog post last week, Zimei Bian talked about how the new mobile app, Pokémon GO, is sweeping the world and how games, and the concept of fun in particular, can be used to engage the public in sustainability efforts.  Also mentioned was Cornell’s eBird project, a web and mobile application that uses crowdsourcing to document the presence and absence of various bird species around the world.  Through this app, birdwatchers can submit data on birds found in their region and can even take a look at hotspots in other regions.  The inexperienced birdwatcher or the casual user who just likes to use the app for kicks can also use Merlin, another mobile app developed by the Cornell Lab of Ornithology, in conjunction with eBird to help identify birds.

Video: A promotion for Merlin Bird ID App (1 minute 54 sec) — Credit to Cornell Lab of Ornithology

But the Cornell Lab of Ornithology didn’t stop there!  Taking advantage of public’s love of games and competition, the lab, in close collaboration with the Cornell Institute for Computational Sustainability,  also developed a treasure-hunting app to go along with eBird called Avicaching, described  in the talk by Yexiang Xue of Cornell University.   The game combines eBird and geocaching to encourage users to search for birds in underrepresented locations by letting you earn a variable number of points for each location visited.  Equipped with a leaderboard that updates in real-time, the game encourages users to visit places with the most ‘points’, helping to reduce eBird’s sampling bias and collect a more accurate distribution of the bird population for scientists to use.  Currently, the game is in the process of development, having only been deployed in two New York counties, but with the initial success seen by the game, it may potentially be expanded into other regions.

Though they were quite memorable for how they built on top of and supported each other, eBird, Merlin, and Avicaching were not the only wonderful projects aimed at selling computational sustainability to public that stuck out to me at the conference.

Another great example of a project that had some focus on public perception included a talk given by David Shmoys of Cornell University on the rebalancing problem found in bike sharing.  Keeping bike racks in balance amidst fluctuating demand is key; a full bike rack will prevent users from stashing their bikes when they’re done and an empty one will deter potential users.  Therefore, bike racks must be managed and optimized to provide an appropriate level of stock at any given time during the day.

A photo of a Citibike station in New York -- credit to Wikimedia user Jim.henderson
A photo of a Citibike station in New York — credit to Wikimedia user Jim.henderson

Similarly, in taxi systems, there is also a demand problem and the issue of user perception.  In his talk on Smart Cities, Bo An of Nanyang Technological University described the peak time dilemma faced by many taxi users in cities: not being able to get a taxi. Because most taxis are priced by distance and because traffic is very slow during peak times, many taxi drivers will simply not work during those hours! The drivers do not believe that driving during peak times is cost effective.  Just like any other user, taxi drivers will attempt to game the system to their advantage, finding the best times to work to maximize their profit.  In order to “sell” the idea of working during peak periods, an incentive like raised fare prices must be brought to the table.  Further detail on this pricing and scheduling scheme can be found here. Its also important that we study possible rebound effects that might result from making “dysfunctional systems more tolerable“.

In a field like computational sustainability, which aims to touch all lives, human and otherwise, now and in the future, by building a better tomorrow, it is especially important that the public is “sold” on what we’re creating, lest the phrase “Everybody’s innovating, nobody’s integrating” becomes applied to us.  Fortunately, as seen by the conference, many researchers are already a step ahead and are proactive.

The projects mentioned here were only some of my favorite ones from the conference that touched on the idea of “selling” sustainability to the public. A complete list of talks hosted at the conference can be found here.

Selina Chen is a Computer Science  undergraduate at Vanderbilt University, with a love for sci-fi novels, superheroes, and art.  Currently, she’s having fun exploring the various ways art can be used to make data beautiful and engaging for the public.The opinions expressed herein are Selina’s and not necessarily those of Cornell University. You can reach Selina at


Science Outreach that Lasts: Scouting

I hiked around Ithaca, NY, the day before the 4th International Conference on Computational Sustainability  at Cornell University, and happened upon a knickknack shop, where I found a cache of 1963 Boy Scout merit badge pamphlets. I bought two that I didn’t have already — Gardening and Bookbinding. Gardening was one of the original merit badges that were introduced in 1911. The Gardening pamphlet, at least the 1963 printing, was written by Professor Paul Work of Cornell University, probably in the 1940s when the material was copyrighted. Professor Work died in 1959, after a distinguished career, which  included authoring The Tomato  —  you’ll see that Professor Work apparently liked to put faces to science, as part of the prefacing material to that book.


Image: The Gardening Merit Badge by Professor Paul Work. Click to enlarge.

I haven’t researched the history yet, but Boy Scout merit badges are my earliest recollection, as a scout myself, of formalized mechanisms of promoting lifelong and project-based learning through badging, and communicating science and technology to the public. Professor Work’s outreach on gardening may seem closer to hobbyist than to scientific material, but there is science outreach in that badge, and among the other 1911 merit badges were those that were clearly science outreach, including Astronomy, Ornithology (later Bird Study), Chemistry, and Electricity. Still others of the originals had sustainability connections as well, to include Conservation, Agriculture, and Forestry.

Eagle Scout Bridge

Image: Bridge in Cumberland Mountain State Park created by Lee Suydam for his Eagle Scout project (2005). Such projects, and there are many, are exemplars of project-based learning, though they are outside the scope of formal school curricula. Click to enlarge. Photo Credit: Douglas H. Fisher

The Boy Scouts of America (BSA) are one of the very first environmental groups in America, and while BSA has been “dragged kicking and screaming” into inclusiveness on some social issues (see Treehugger article), they have been environmentalists consistently. The current crop of sustainability-relevant merit badges are many: Animal Science; Architecture; Bird Study; Composite Materials; Energy; Environmental Science; Fish and Wildlife Management; Forestry; Geology; Insect Study; Landscape Architecture; Mammal Study; Mining in Society; Nature; Nuclear Science; Oceanography; Plant Science; Reptile and Amphibian Study; Soil and Water Conservation; and Sustainability. Moreover, among the required badges for Eagle Scout is either Environmental Science or Sustainability (choose at least one).  A summary history of all merit badges, past and present, is an interesting read, …, for those interested (like me)!

After CompSust-2016, I went to Nashville’s Scout shop and picked up many of the study pamphlets for sustainability-related merit badges,  and was gratified to find attention to climate change in the most recent Sustainability merit badge (instituted 2013), and as importantly, global warming, climate change, and greenhouse effects have found their way into the study pamphlets of older merit badges like Chemistry, Weather, Environmental Science, and others. This article in Treehugger points to exactly the satisfaction and mild surprise that I found in the BSA environmental record since I was last active.

BSA has a long history of technology-relevant merit badges too (e.g Machinery, 1911 – 1995). In “my day” there were badges on Computers (1967-2014), Electronics (1963 – ), Engineering (1967 – ), which has morphed and grown to include Digital Technology; Robotics; Programming; GeocachingGame Design; Entrepreneurship; and Graphic Arts.

This brings me to a  goal of infusing computational sustainability  into the BSA merit badge system. Our NSF-funded network, CompSustNet,  aspires to introduce and nurture outreach in K-12 that reaches far and wide, and that lasts well beyond the network’s NSF funding period. Scouting has a long and proven history of science and engineering outreach (as well as Arts and Humanities outreach — just look at the merit badge list), so its no surprise that we are investigating the outreach possibilities with scouting. While I have focused on BSA, which is integral to my personal story, I am learning about Girl Scouts of the USA (GSUSA) and their badging system, with goals for outreach in GSUSA as well.

Web searches with keywords such as “NSF” (or “National Science Foundation”), “Boy Scouts” and “merit badge” show that there are NSF grants with broader impacts plans that include activities with scouting, and merit badge workshops and study groups (e.g., “CAREER: Computational Modeling of Microstructure Evolution during Vapor Deposition).  Additional poking around finds that there are museums around the country that work with scouts as part of the museum’s disciplinary outreach (e.g., Nashville’s Adventure Science Museum). Museums and other institutions can have their own (digital) badging systems, and we are currently designing the desiderata, requirements, and graphic designs of CompSust badges.

We can aspire to create BSA and GSUSA badges on Computational Sustainability, but in the near term, our focus is on workshop materials that scouts and their mentors can use to integrate computing into satisfaction of sustainability-themed badge requirements, and to integrate sustainability into computing-themed badges.

One “secret formula” of BSA, at least in my experience, is that the “library” research involved in merit badges, ecology-themed and otherwise, is side by side with in-the-world activities (including merit badges and Eagle projects) in Backpacking; Cooking; Gardening; Scuba Diving; Search and Rescue; Climbing; Fishing; Citizenship in the Community, Nation, and the World; and so many others — that amalgamates interests in nature, citizenship, science, and humanities — at least it did in me.


Image: Glendora Explorer Post 494 scouts and leaders hiking Mount Whitney (circa 1974). Click to enlarge. Photo Credit: Douglas H. Fisher

Infusing computational sustainability into scouts (and other badging systems and K-12 institutions) appears very promising. There will be other outreach efforts by CompSustNet that are intended as long-lasting and consequential as well — we will be posting these activities to this blog,

Thanks to Professor Paul Work too, for being a pioneer in communicating science to the public. It was serendipity that I discovered him, but serendipity that was made more probable by a curiosity about and appreciation for the place I was in.

Douglas H. Fisher is CompSustNet’s Director of Outreach, Education, Diversity, and Synthesis. The opinions expressed herein are Doug’s and not necessarily those of Cornell University. Contact Doug at 

Computational Sustainability for Everyone: Untapping the Potential of Games, As Told by Pokémon GO

This is a post from Zimei Bian. See her bio at the bottom.

If you haven’t been living under a rock for the past two weeks, you’ve probably heard about the new cultural phenomenon sweeping the world–at least in countries where the game has been officially (or unofficially) released — that is Pokémon GO.

Pokemon logo large

(Pokémon Go logo, property of The Pokémon Company)

So what is Pokémon GO, and why are we talking about it on a Computational Sustainability blog? In brief, Pokémon GO is a location-based augmented reality (AR) game on mobile devices that allows players to capture and train virtual creatures “encountered” in the real world from the Pokémon franchise. Players can “hatch” Pokémon eggs by walking certain distances and battle other players for control of Pokémon “gyms” at various landmarks. With the exception of completely optional in-game microtransactions, the application is free to download and play with the help of the built-in GPS and camera capabilities available on most modern iOS and Android handheld devices.


(Screenshots of Pokémon GO gameplay. Image credit: Zimei Bian)

In a little over a week since its release, Pokémon GO has not only surpassed Twitter in terms of active daily users, but has made significant contributions everywhere from increasing voter registration to improving users’ mental health. Looking at these reports, the question arises: can games like Pokémon GO play a part in sustainability efforts? The answer, as it turns out, is that it already has.

In addition to motivating people to explore National Parks, helping communities fall in love with their cities, and encouraging players to pick up trash in public spaces, Pokémon GO has created ample opportunities for users to encounter real life plants and animals as they track down fictional Pokémon (many of which are based on real animals). Scientists on Twitter have caught on to this phenomenon and have created citizen science initiatives under the hashtags #PokeBlitz and #PokemonIRL to help Pokémon GO users identify and learn about the wildlife they come across on their adventures. Judging by the influx of posts with these hashtags, it appears that many users are interested in more than just virtual creatures. But how can we further harness this innate curiosity for sustainability? And what role can gaming play in our efforts?


(#PokeBlitz and #PokemonIRL tweets on Twitter)

Two weeks ago, I attended the 4th International Conference on Computational Sustainability at Cornell University, where Computational Sustainability experts and enthusiasts from all over the world gathered to share their research. Many of the talks were centered around wildlife conservation and the potential impact of citizen science in these sustainability efforts. One of my personal favorite projects in this category was eBird, a highly successful CompSustNet project launched in 2002 by the Cornell Lab of Ornithology and National Audubon Society that provides an intuitive web application for bird watchers across the globe to report observations and access information about their personal records as well as interactive visualizations of all collected data. Despite the popularity of the project in the global birding community, one of the points of discussion that arose around eBird and citizen science projects in general (iNaturalist, Project Noah, etc.) was how to attract and retain users that are not already experts–or even prepossess some particularly strong interest–in the field. In this respect, I believe that we can learn a lot from Pokémon GO, an application that has somehow managed to attract an abundance of users from all walks of life — many of whom were not fans of Pokémon (or even exercise, in many cases) in the past.

In her previous post on this blog, Selina Chen wrote about the importance of providing ways for the average citizen to feel that they can be involved in and make substantial contributions to sustainability efforts. One way to do this, as mentioned in the aforementioned post, is by bringing the issues home and highlighting their impact on a local scale. The widespread appeal of Pokémon GO suggests another potential method: by making it so that exploring and contributing to sustainability causes is fun. While this may not seem like a particularly sophisticated suggestion, research shows that having fun motivates further pursuits of knowledge and increases the likelihood of meaningful connections being made during learning in both children and adults. And what does it say about the power of fun that in just a little over a week from the game’s launch, Pokémon GO players from all over the world have already banded together to contribute observations to extensive crowdsourced Pokémon species-sighting maps–an endeavor eerily similar to what wildlife conservation scientists have been encouraging people to do for years? The Fun Theory, an initiative of Volkswagen, is dedicated entirely to the idea that “something as simple as fun is the easiest way to change people’s behaviour for the better.” One illustration of this initiative’s many incredibly simple but effective ideas can be viewed below:

(“Bottle Bank Arcade”, one of many ideas from Volkswagen’s The Fun Theory initiative.)

Winston Churchill once said, “I am always ready to learn although I do not always like being taught.” The reason that this quote has survived for so long after its originator’s death is because we can all relate to it. The same could be said about keeping our New Year’s resolutions and practicing socially-and-environmentally responsible behavior–we all want to do it, but it’s often inconvenient and much easier to just toss to the side. How can we motivate the average person who is undoubtedly already busy with their own lives to learn about, care about, and contribute to solving global sustainability problems? With all the conversation around the wildly successful Pokémon GO right now, there really is no better time to start thinking seriously about how the Computational Sustainability field can utilize games and emerging gaming technologies to engage and motivate a wider demographic to start taking part in our sustainability efforts.

Who knows? We may even have some fun in the process.

Zimei Bian is a CS undergrad at Vanderbilt University with a special passion for interactive storytelling and using tech for social good. In her spare time, she enjoys internet cat pictures and story-driven video games. The opinions expressed herein are Zimei’s and not necessarily those of Cornell University. You can reach Zimei at

Professional Development in Computational Sustainability

The 4th International Conference on Computational Sustainability (CompSust-16) was last week at Cornell University in Ithaca NY. It was a great conference, and videos of all presentations will be online soon. We will be using many of these online presentations as jumping off points to talk about the individual CompSust network projects, so stay tuned to this blog.

My own presentation on the broader impact plans for CompSustNet will be online as well, and my slides are available now.

Naturally, I will be addressing themes found in my talk from time to time. In this post, I briefly highlight professional development (e.g., bullet points at the bottom slides 6 and 13) and the creation of a new LinkedIn group on Computational Sustainability, with a mission as follows.

“Computational sustainability concerns the application of computing to challenges of environmental and societal sustainability, and the research and development required for such applications.”

“This group investigates, reports, and discusses career trajectories in computational sustainability in industry, academia, government, and non-profit sectors. The membership includes professionals, faculty, teachers, students (at all levels) in all the computing and sustainability sciences, as well as general educators, and other interested citizens.”

“All areas of computing — including artificial intelligence, machine learning, database, hardware and operating systems, mobile computing, robotics, multi-agent systems, social computing, visualization, algorithm analysis — have applicability to sustainability. The sustainability areas of importance are vast, including wildlife conservation, climate change mitigation and adaptation, urban design and traffic, disaster management, energy, agriculture, and poverty response.”

Three conversations at CompSust-2016 motivated this addition to CompSustNet’s social media outlets, which also includes Facebook and Twitter.

After my presentation, two undergraduate computer science majors who were attending the conference, suggested a LinkedIn group to learn and network about computational sustainability career opportunities and trajectories, to include the importance of internships and course selection. In fact, they thought that LinkedIn was the most relevant and important of all the social networking sites for purposes of building community.

A PhD student, and then another, expressed growing interests, as graduation approached, in learning more about computational sustainability career opportunities in industry, and our discussion (with Carla too) also raised the potential for government and non-profit opportunities. Clearly there are academic positions in computational sustainability too, which was well illustrated at the conference by the involvement of many new faculty in the area.

Finally, at an evening meeting of conference leadership, the importance of developing and demonstrating career trajectories in computational sustainability was discussed at length.

The LinkedIn forum will followup with thoughts that flow from the CompSust-2016 conference on career opportunities. We also plan to create videos from professionals who have followed a computational sustainability career path, and to make such videos a regular contribution of the Linked In group.

Douglas H. Fisher is CompSustNet’s Director of Outreach, Education, Diversity, and Synthesis. The opinions expressed herein are Doug’s and not necessarily those of Cornell University. Contact Doug at 

Surveys by CompSustNet

CompSustNet is collecting information and feedback in two forms. Please fill one or both out, as you deem appropriate.

Census of CompSustNet Participants: This questionnaire is to collect basic data on participants, broadly construed, of CompSustNet, as well as to collect information more broadly about those working on (or at least interested in) problems of computational sustainability.  If you want to be added to the CompSustNet mailing list, please fill out this form. All questions are optional, but we hope to get your name and affiliation at a minimum.

Feedback on OEDS at CompSust-16: We welcome feedback on issues related to Outreach, Education, Diversity, and Synthesis (OEDS) for CompSustNet. All questions are optional. Questions 4-8 follow the broad areas outlined in the OEDS presentation at CompSust-2016. The final question is an opportunity for less structured comments. Feel free to email any other thoughts on OEDS issues at any time to

Douglas H. Fisher is CompSustNet’s Director of Outreach, Education, Diversity, and Synthesis. The opinions expressed herein are Doug’s and not necessarily those of Cornell University. Contact Doug at