![]() Below is a list of geospatial-intelligence technologies we are seeking:Īlthough we are broadly seeking unique geospatial-intelligence technologies, some areas of specific interest include: The program is searching for best-in-breed businesses developing novel dual-use technologies (products that have both commercial and military uses), which will grow the number of geospatial solutions available to NGA and its customers. What types of companies should apply to the NGA Accelerator? MACHINE-LEARNING TRAINING DATA: NGA scientists need analysis-ready, structured, and validated data sets to support dynamic algorithms and model development that is repeatable and auditable for new and existing use cases. MULTI-DIMENSIONAL DATA MANAGEMENT: NGA systems engineers need to organize and store geospatial information in data models that reflect temporal currency in order to produce more comprehensive intelligence products.ĪUTOMATED OBJECT VALIDATION: NGA analysts need a way to automatically distinguish between multiple observations of the same object and observations of similar objects to validate the accuracy of imagery-based observations.ĪUTOMATED CONDITIONING SERVICE: NGA needs the ability to automate data conditioning for easy and accurate metadata tagging so that data is discoverable and accessible by the enterprise.ĮNTITY RELATIONSHIP MODELING: NGA analysts need to conflate the ontologies, schema, and knowledge graphs of multiple non-pixel datasets and model the relationships between them. MULTI-DOMAIN DATA AGGREGATION: NGA analysts need to rapidly aggregate diverse data types and schemas from sources across multiple domains to quickly extract intelligence at scale. There is an increasing demand for timely and agile integration of new data management technologies.ĭATA INTEGRATION: NGA analysts need the ability to rapidly discover and integrate diverse data types and multiple types of pixel- and non-pixel-based data from multiple sources to discover and characterize relevant patterns. Data vary in source, quality, reliability, format, and transmission method and include diverse collection methods, such as space-based satellites, airborne platforms, maritime vessels, foreign governments, and open sources. The rapid advancement of technology and the increasing density of interconnected devices and sensors has led to an explosion of geospatial data from sources with diverse organizational schemas and security requirements. OBJECT CLASSIFICATION: NGA analysts need to categorize detected objects at a broad level of organization or feature and further classify within that category.įEATURE CHARACTERIZATION: NGA analysts need to automatically characterize the attribution information of extracted foundation geospatial data in order to build upon the knowledge recorded from previous recognition, identification, and classification of objects with the highest level of specificity. OBJECT DETECTION: NGA analysts need to detect objects and changes in an image with little or no description in order to assess movement or changes to features.įEATURE EXTRACTION: NGA analysts need to extract geospatial data feature geometries and their attributes with the highest measure of accuracy and speed from various sources in order to reduce labor-intensive manual extraction hours. ![]() Under the DoD’s ethical principles of AI, machine assisted processes will be responsible in their development, deployment, and use equitable with minimal unintended bias transparent and auditable reliable through extensive testing and governable throughout the lifecycle.ĪCTIVITY CONTEXTUALIZATION: NGA analysts need to extract and correlate tracks of activity from non-imaging sources, such as radio frequency and automatic tracking system data, with more traditional geospatial data, such as imagery, to contextualize changes in an area of interest. ![]() NGA’s needs for high confidence in AI and understanding of the generated output under varying conditions are supported by the larger Department of Defense (DoD) and intelligence communities. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |