Project Description
Inspired by recent advancements in geospace observing capabilities and the opportunities of Big Data, our research group developed and deployed an open-source Python software and associated web-applications for Assimilative Mapping of Geospace Observations (AMGeO) that are fully interoperable with established community data resources supported by NSF and NASA programs. This DLA project will capitalize on this AMGeO data science cyberinfrastructure to produce training data sets and develop a predictive model of geospace conditions using machine learning and artificial intelligence analysis techniques. The DLA student will learn how to use AMGeO, and apply recurrent neural networks and hyperparameter optimization techniques to develop the model by building on our earlier work. The DLA student will also participate in various aspects of the full-stack AMGeO cyberinfrastructure development, including but not limited to, microservice architectures, docker infrastructures, auxiliary artificial intelligence and machine learning analysis tools, and visualization tools.
Special Requirement
Strong programming skills. Be proficient with Python programming language and familiar with linux operating system. Be available to work at least in one 5-hour block with Aerospace Engineering Sciences staff on a weekly basis. Familiarities with Python libraries such as Numpy, Xarray, Matplotlib, Tensorflow as well as Flask and Docker are desirable. We are looking for a quantified and self-motivated student with interests and experience in artificial intelligence and machine learning techniques.
Contact
- Tomoko Matsuo (faculty)
- Willem Mirkovich (professional research assistant)