UDiscoverIt: Incentivizing Citizen Science Discovery for a Sustainable World

Saturday, February 13, 2016: 8:00 AM-9:30 AM
Marriott Balcony A (Marriott Wardman Park)
Carla P. Gomes, Cornell University, Ithaca, NY
Gomes will talk about UDiscoverIt, a program of Cornell’s Institute for Computational Sustainability that seeks to accelerate scientific discovery by integrating citizen science data and crowdsourced information into advanced computational models and algorithms. Gomes will describe several projects, ranging from species conservation to poverty mitigation in Africa and materials discovery for renewable energy.

In one example, Gomes will describe how novel computational statistical and learning models, combining eBird data with environmental data, predict bird species occurrence across broad spatial and temporal scales. The models are capable of discovering the patterns of occurrence of birds with the sparse, noisy data collected by citizen scientists, by relating environmental features to observed bird occurrences and making predictions at un-sampled locations and times. eBird is a citizen science program of the Cornell Lab of Ornithology that has collected millions of bird observations.  Gomes will also talk about Avicaching, an exciting game to incentivize birders to submit bird observations from habitats that are generally under-sampled by normal eBirding,

In another example, Gomes will talk about GrazeIt, a project that aims to improve rangeland and forage maps in Africa using vegetation images and surveys submitted by herders. A pilot project in Kenya collected over 100,000 images and surveys in just a few months.

 Gomes will also highlight the power of combining state-of-the-art computational models with insights from citizen scientist to accelerate materials discovery for sustainable energy materials, such as fuel cells and solar fuels. In one of the projects, the goal is to identify the crystal structure of materials based on x-ray diffraction data. Gomes will talk about an approach that combines the pattern recognition abilities of citizen scientists to interpret the x-ray diffraction data with advanced computational techniques. This approach, merging human and artificial intelligence, outperforms those produced solely by humans or computers.