calypsoai logo

CalypsoAI completes US Air Force SBIR Phase 1

May 12, 2020


CalypsoAI recently completed a SBIR Phase 1 project with the US Air Force to accelerate the use of secure AI across the service. CalypsoAI’s efforts focused on three primary lines of effort, all of which leverage CalypsoAI’s commercially available model risk management platform, Vespr, to meet USAF mission needs. The three lines of effort include focusing on the three primary challenges faced by the USAF in the service’s acceleration of AI. These include:

  1. Lack of Transparency – As the USAF integrates AI across core mission functions, more non-data scientists are needing to understand and access models. But teams lack the tools to evaluate model quality and cannot monitor models for ongoing performance.
  2. Changing AI Security Requirements – As attacks on AI rapidly emerge, especially in national security uses, AI users are rarely abreast of AI vulnerabilities and best-practice AI security techniques. AI developers and users require tools to test the security and robustness of their models against adversarial manipulation.
  3. Lack of Capability – The national security industry struggles to attract and retain top data science talent which leads to unverifiable model quality and unknown testing procedures. Meanwhile, organizations with data scientists building AI tools lack model validation and & verification capabilities.

Through our Phase 1 efforts, Calypso demonstrated how Vespr, combined with CalypsoAI’s advanced research and development team, can help accelerate USAF AI efforts in a secure, transparent manner while enabling Airmen to build models at the edge.

Subscribe to our newsletter

Stay up-to-date on our latest developments
with our monthly newsletter.