The SERVIR Network, working within five regional hubs around the globe, continues to address environmental issues by enabling the access, implementation, and application of Earth observations. Over the past two years, the SERVIR Network has made an effort to build capacity in utilizing Machine Learning (ML), Deep Learning (DL), and Artificial Intelligence (AI) approaches to address these environmental challenges.
The TensorFlow Working Group (TFWG) has been a spin-off of this initiative. The TFWG serves as a capacity-building forum for members from the SERVIR hubs, NASA Applied Sciences Team members, private sector, and university researchers to collaborate and grow abilities to directly apply ML, DL, and AI techniques.
This presentation highlighted capacity-building efforts, collaboration opportunities, and SERVIR's vision for applied Machine Learning service development.
Speaker: Tim Mayer (NASA SERVIR Hindu Kush Himalaya Regional Science Coordination Lead)
Tim Mayer holds a BS in Biology and a MS in Environmental Science. His primary research interests include machine learning applications of remote sensing focusing on landscape ecology and land cover/land use analysis. With the NASA SERVIR Science Coordination Office, Tim contributes to the co-development and generation of applied remote sensing services to address an array of environmental challenges. Additionally, he works to enable the use of Earth observations through building capacity efforts.
If you're interested in presenting your research or perhaps you have some questions, contact timothy.j.mayer@nasa.gov to schedule joining one of TFWG meetings.
This answers my question.
Thank you for your feedback
The meetings happen biweekly on Wednesdays at 10:00 am Central Time.
This answers my question.
Thank you for your feedback
It is important for us to share our solutions and progress so we try to make everything open source as often as possible.
This answers my question.
Thank you for your feedback
Yes, some algorithms are certainly changing fast, but we are comfortable with the degree of how much we don't know about the inner workings of the tools. Since the primary goal is to provide geospatial services, it pays off to use Google's black box tools as they provide a lot of value to us. In the future though we do plan to explore the more open source alternatives.
This answers my question.
Thank you for your feedback
We try to pair complete beginners that wish to learn and become experts with our own experts from TFWG. Usually this guidance is for a week and afterwards you keep learning by participating in the technical meetings and asking questions.
This answers my question.
Thank you for your feedback
Yes it is, although we had to supplement the default GEE SAR data with our own.
This answers my question.
Thank you for your feedback
One of the biggest lessons learned for us is to incorporate experts at every turn. This helps to avoid getting stuck in a project, and almost always brings new technologies and perspectives to the table. It also helps with keeping everyone motivated and excited to work.
