The National Geospatial-Intelligence Agency (NGA) is embracing the power of artificial intelligence (AI) and machine learning (ML) to enhance data analysis in record time. Leveraging AI and ML, the agency aims to improve the processing, management, and interpretation of vast amounts of data.
One of the primary applications of AI and ML at NGA is the identification and analysis of satellite imagery, with ongoing efforts towards developing improved data management techniques. However, NGA acknowledges that the human factor remains crucial in addressing challenges such as bias in data collection, managing output drift, and maintaining security amidst technological advancements.
Natasha Krell, a computer vision and machine learning scientist at NGA, emphasizes the importance of AI in enabling individuals to perform their jobs more efficiently. Drawing an analogy to a word document processor, Krell explains that AI functions as a tool to assist in various tasks rather than replace human input entirely.
The adoption of AI and ML tools by NGA allows for better detection and classification of objects within the deluge of satellite data, enhancing insights derived from imagery analysis. NGA examines both on-premise and cloud-based opportunities for storage, infrastructure, and compute resources.
To further advance AI and ML within the agency, NGA combines its large multimodal models for electro-optical, synthetic aperture radar (SAR), and thermal imagery. This integration represents the cutting edge of AI utilization according to Krell.
NGA also collaborates with government partners, sharing output models, data, and insights as part of their commitment to supporting humanitarian assistance and disaster relief missions.
While AI and ML show immense potential, NGA faces certain challenges in their implementation. Training the workforce, managing the volume and velocity of data, and mitigating bias are among the top concerns. Robust data management systems are crucial for AI, given its substantial data requirements, and the agency remains mindful of bias stemming from datasets, models, and algorithms.
Security risks associated with AI and ML tools are also a significant consideration for NGA. The agency emphasizes the importance of rigorous verification and validation of algorithms and models, ensuring that their implementation aligns with guidance provided by the chief data and AI officer and other relevant entities.
NGA recognizes the role of tools like ChatGPT in increasing public and workforce understanding of AI and ML. Natural language processing, in particular, is gaining traction in academia and industry, offering exciting advancements in large multimodal models that combine text and imagery for more nuanced insights.
In conclusion, NGA’s adoption of AI and ML is revolutionizing their ability to process and analyze vast amounts of data. By leveraging these tools, the agency aims to enhance satellite imagery analysis, address humanitarian challenges, and improve their overall efficiency. However, NGA also acknowledges the importance of human expertise in tackling bias, maintaining security, and maximizing the potential of these technologies.
