Scaling Machine Learning for Remote Sensing using Cloud Computing
Learning outcomes
• Setting up a cloud computing environment
• Preparing ML data training data
• Developing, evaluating and deploying ML models
• Performing model inference on new data (put model in production)
Lecture content
Machine Learning (ML) is not just about finding the right algorithm and creating a model. It is about the entire end to end lifecycle. The ML lifecycle includes four phases: problem definition, data collection and analysis, model development and evaluation, and deployment to a production system. The availability of open Earth science data offers immense potential for ML as evident from numerous research publications lately. However, many of these publications are not ending up as production applications mainly because the data scientists who develop the ML models are now expected to deploy and scale the models in production. Since more and more remote sensing data are being made available in cloud computing environments, scaling and deploying the ML models can be streamlined using cloud backed services.
This lecture will introduce ML lifecycle to the participants and demonstrate end-to-end remote sensing ML application from data preparation to deployment using a cloud computing environment.
The Instructors
Dr. Manil Maskey
Bio
Manil Maskey received the Ph.D. degree in computer science from the University of Alabama in Huntsville, Huntsville, AL, USA. He is a Senior Research Scientist with the National Aeronautics and Space Administration (NASA), Marshall Space Flight Center, Huntsville. He also leads the Advanced Concepts team, within the Inter Agency Implementation and Advanced Concepts. His research interests include computer vision, visualization, knowledge discovery, cloud computing, and data analytics. Dr. Maskey's career spans over 20 years in academia, industry, and government. Currently he chairs the IEEE Geoscience and Remote Sensing Society and Earth Science Informatics Technical Committee, and leads the machine learning activities for the NASA Earth Science Data Systems program.
Iksha Gurung
Bio
Iksha Gurung is a Computer Scientist working with University of Alabama in Huntsville, supporting National Aeronautics and Space Administration Inter-Agency Implementation of Advanced Concepts Team (NASA-IMPACT). He leads the development and machine learning team in NASA-IMPACT.
Shubhankar Gahlot
Bio
Shubhankar Gahlot received his MS in Data Science from Illinois Institute of Technology Chicago, USA and BS in Industrial Design from Indian Institute of Technology Guwahati, India. He is a Research Scientist at The University of Alabama Huntsville for NASA IMPACT team. He has more than 2 years experience in ML and ML Ops at scale at Oak Ridge National Lab, USA and produced multiple publications on how to benchmark and scale ML on supercomputers. Prior to that he has worked in the software industry as a design lead and architect. He is passionate about machine learning applications, its benchmarking and reproducibility and has contributed to MLPerf (now MLCommons), an organization that builds fair and useful benchmarks for measuring training and inference performance of ML hardware, software, and services.
Shubhankar Gahlot
Bio
Shubhankar Gahlot received his MS in Data Science from Illinois Institute of Technology Chicago, USA and BS in Industrial Design from Indian Institute of Technology Guwahati, India. He is a Research Scientist at The University of Alabama Huntsville for NASA IMPACT team. He has more than 2 years experience in ML and ML Ops at scale at Oak Ridge National Lab, USA and produced multiple publications on how to benchmark and scale ML on supercomputers. Prior to that he has worked in the software industry as a design lead and architect. He is passionate about machine learning applications, its benchmarking and reproducibility and has contributed to MLPerf (now MLCommons), an organization that builds fair and useful benchmarks for measuring training and inference performance of ML hardware, software, and services.
Drew Bollinger
Bio
Drew Bollinger leads the Labs Team at Development Seed. Drew is a data analyst, software developer, and machine learning engineer, with experience building geo-interfaces and running advanced statistical and spatial analysis on open data sets. He has delivered several impactful workshops at Satsummit, IGARSS, and most recently at the Africa Geospatial Data and Internet Conference in Accra, Ghana.