Geospatial Intelligence & AI/ML During a Pandemic

The recent global working climate, altered by COVID-19, has driven considerable change in how the geospatial community delivers support and new capabilities for artificial intelligence and machine learning (AI/ML). The pandemic forced analysts to work from home without the advanced infrastructure required for ML, but the high demands of mission requirements remained. Therefore, the geospatial community had to adapt. New technologies are accommodating workflow gaps, existing technologies are being leveraged in new and innovative ways, and from an operational and cultural perspective, teams are embracing new approaches to their work.

During the USGIF GEOINT Community Forum Nov. 16-18, panelists from USGIF’s Machine Learning and Artificial Intelligence Working Group (MLAI WG) discussed the impact of the pandemic, assessed the accomplishments and challenges of the GEOINT community, and forecasted what to expect moving forward.

Historically in classified environments, such as that at the National Geospatial-Intelligence Agency (NGA), there is a limited need for AI/ML solutions created in the unclassified space. However, due to the pandemic, working from home has left many without the infrastructure to keep up with AI/ML delivery. Therefore, a large investment in higher-performance data processing environments was recently made for AI/ML.

Maxar is just one of the many organizations to work with NGA during this pandemic. Maxar helped NGA expand the adoption of telework-friendly, collaborative mapping tools, and access to commercial satellite imagery.

“Users with a browser and internet service can leverage a combination of [unclassified tools like] AI/ML models, automation tools, and human expertise to improve the volume and quality of foundational maps throughout the world,” said Todd Bacastow, senior director for strategic growth and emerging technologies at Maxar.

Also, according to Wynn Thane, associate director, Guidehouse, many developers of these tools and models have created them on the unclassified side and now the key is to integrate them to the ‘high side.’ This “develop low, deploy high” mentality provided a strong foundation for continuing development of AI/ML initiatives in the wake of COVID-19.

BAE Systems is one of the organizations that has adopted this “develop low, deploy high” mentality, specifically on the AI side.

“We’ve taken some of our algorithms and worked with teams we’ve partnered with, like AWS, to move them to the intelligence community (IC) marketplace. And that is an easy path for us to take these algorithms that are at a low technology readiness level right now and put them in the hand of developers in the IC,” said Don Widener, Ph.D., technical director, BAE Systems.

Additionally, BAE Systems telework environments have allowed for a major upskill training push with Robotic Process Automation (RPA). RPA enables analysts with a limited programming foundation to build automation models or “bots” to address repetitive tasks. In the recent USGIF white paper, “Geospatial Intelligence & AI/ML Progress During a Pandemic,” Widener reports that geospatial open-source intelligence analysts have been utilizing RPA for data collection and processing, further driving efficiency in AI/ML delivery.

While COVID-19 caused many initial concerns, the geospatial community adapted and continues to progress with AI/ML initiatives and efforts. Adapting to the workforce impacts of COVID-19 has had some benefits to AI/ML delivery.

David Gleason, a solutions architect with Amazon Web Services (AWS), reports that AWS has seen an increased demand for unclassified AI/ML environments, training on latest technologies, and management of services for AI/ML for data labeling, training, and deploying solutions.

“There was a really big emphasis on building low-side environments to maintain and be productive for our workforce. That has started to empower folks to think differently about being builders and developing the right toolsets to be productive and mobile on the classified environments—we have seen this trend increase during the pandemic,” said Gleason. The path forward can certainly build from lessons learned to overcome the challenges and maximize the benefits of the COVID-19 work environment.

• USGIF has released a white paper on a topic impacting the entire GEOINT community: “Geospatial Intelligence & AI/ML Progress During a Pandemic.” Read the entire white paper, sponsored by BAE Systems, Guidehouse, and Maxar.