You are here

SDS Lab PhD students Kate Duffy and Udit Bhatia excel at NASA and in Big Data

September 26, 2018

SDS Lab PhD student Kate Duffy completed a successful summer internship at NASA Ames. In her own words: "This summer at NASA Ames I worked to develop a deep learning approach to atmospheric correction, the process in which top-of-atmosphere satellite images are processed to remove the effects of gases and aerosols in the atmosphere. The surface reflectance produced is used to develop useful higher-level land products, such as crop classification and burned area. A model is trained to learn a mapping from top-of-atmosphere images from a geostationary satellite to surface reflectance retrieved from MODIS. This approach relies on the ability of neural networks to complete a style transfer-like task, with validation using traditionally derived geostationary surface reflectance. This work represents a step toward retrieving a higher temporal resolution surface reflectance product than is available from polar satellites, with lessened need for atmospheric data collection and lighter computational demand. Additionally, I had the opportunity to tour NASA’s Pleiades Supercomputer facility and collaborated on a white paper report exploring cloud computing options in the context of machine learning and comparing them to NASA’s high-performance computing capabilities." Duffy was mentored by NASA scientists Sangram Ganguly and Rama Nemani. While at NASA, she also got an opportunity to work with recent SDS Lab alumnus Thomas J. ("TJ") Vandal, who recently joined NASA as a scientist after completing his PhD.  

SDS Lab PhD student Udit Bhatia, worked with his Northeastern adviser and with scientists at US DOE's Pacific Northwest National Laboratory to help win a Lab seed grant led by PNNL Senior Scientist and Team Lead Samrat (Sam) Chatterjee. Bhatia is the first author of a paper titled "Aviation Transportation, Cyber Threats, and Network-of-Networks: Conceptual Framing and Modeling Perspectives for Translating Theory to Practice" accepted at the 2018 IEEE International Symposium on Technologies for Homeland Security. Bhatia was interviewed (click for journal article) by the journal Big Data about the textbook book titled "Critical Infrastructures Resilience: Policy and Engineering Principles", which he co-authored with two Northeastern faculty. 

Duffy and Bhatia have one peer-reviewed paper each at the upcoming Eighth Workshop on Data Mining in Earth System Science (DMESS 2018), which will be held in conjunction with the 2018 IEEE International Conference on Data Mining in Singapore on November 2018.