Skip to end of banner Go to start of banner

Artificial Intelligence/Machine Learning

Skip to end of metadata

Created by Hsu, Leslie, last modified on Nov 08, 2021

Go to start of metadata

Description: A group to share information related to Artificial Intelligence and Machine Learning.

Points Of Contact: JC Nelson (jcnelson@usgs.gov), Pete Doucette (pdoucette@usgs.gov)

Meetings: 2nd Tuesdays of the month from 3-4pm ET

To join: Add your email to https://listserv.usgs.gov/mailman/listinfo/cdi-aiml

Go to: Meeting Notes Forum

NEW USGS AI/ML Inventory

Please share the inventory link with anyone in USGS that is working on an AI related project. The CDI AI/ML working group will be sharing this information with the 21st Century Science/EarthMap working group, as well as with the Regions and Mission Areas. For the initial inventory of projects, see FY19 AI/ML Inventory Spreadsheet.

Purpose

The purpose of the group is to discover and share AI/ML work being done at the USGS, offer learning opportunities, learn what needs exist across the bureau, and create connections between USGS employees.

Using the themes generate during the Nov CDI call here is a word cloud highlighting people's interests in AI/ML

Meeting Schedule

Next meeting, November 9, 2021

Title: Migration of NLCD and RCMAP to Neural Net classification Systems

Description:

The National Land Cover Database (NLCD) and Rangeland Condition, Monitoring, Assessment, and Projection (RCMAP) will give an overview of progress migrating to cloud processing and Neural Net classification systems for National Scale Production. This progress includes comparison of decision tree classifiers and analysis of modeling inputs to increase accuracy and decrease post processing manipulation.

Presenters:

Patrick Danielson
Technical Support Services Contract (TSSC)
Patrick leads the development and operational components for NLCD, which includes image compositing, AI-Deep Learning, and scaling processes to run on the High Performance Computer (HPC) and in the cloud on AWS.

Kory Postma
Technical Support Services Contract (TSSC)
Kory Postma worked with Boeing on the SETE (Software Engineering Tools and Environments) program and with JHU/APL (The Johns Hopkins University Applied Physics Laboratory) on the DWC (Distributed Weapons Coordination) program before joining USGS EROS in 2009 and is currently the MAA (Mission Acquisition Analyst) for Landsat 8 and Landsat 9 and is responsible for modifying acquisition strategies to support the LTAP (Long-Term Acquisition Plan). Outside of work, Kory has become an expert in game audio programming and was one of the founders of a video game studio that has produced multiple #1 top-selling video games on Steam and has published chapters in technical books related to that field.

Jon Dewitz
NLCD Production Manager
Jon has created NLCD products for twenty years and has led NLCD production for the last 5. He has a background in computer Science, Physics, and Geography.

Matthew Rigge
Matthew is an ecologist working for the U.S Geological Survey at the USGS Earth Resources Observations and Science Center in Sioux Falls, SD. His focus is quantifying the current and historical condition of rangelands across the western United States. These data fuel investigations into the drivers of change in rangeland systems, including climate change, interannual weather variation, and land management practices.