On May 3, the National Geospatial-Intelligence Agency (NGA) released the updated version of Topic 9 for the BIG-R BAA: Identifying Adversarial Machine Target Generated Overhead Imagery (IAGO). Abstracts are due by 5:00 p.m. Eastern on May 10, and proposals are due by 5:00 p.m. Eastern on June 21.
NGA Research requires assistance to:
- Investigate and develop methods and algorithms, applicable to electro-optical (EO) imagery, for the detection, explanation, and mitigation of adversarial noise attacks on overhead classification and detection models. Adversarial noise is also known as adversarial examples or evasion attacks.
- Investigate and develop methods and algorithms, applicable to electro-optical (EO) imagery, for the detection, explanation, and mitigation of adversarial patch attacks on overhead classification and detection models.
IAGO pursues two objectives: defense and mitigation of adversarial noise and adversarial patches on EO overhead data. These objectives require both EO imagery processing and data science expertise.
The United States Government is seeking proposals for research on detecting and mitigating adversarial noise and adversarial patches in EO overhead imagery. Adversarial noise and patches are a threat to machine learning systems that can cause them to misclassify or malfunction. This research is intended to improve the security of machine learning capabilities used on EO overhead imagery applications, such as satellite and aerial imaging.
The Government plans to make two awards, but may make one award, or none. Contract award under this topic is subject to availability of funds.
Review the NGA IAGO BIG-R BAA topic.
Source: SAM
The right opportunity can be worth millions. Don’t miss out on the latest IC-focused RFI, BAA, industry day, and RFP information – subscribe to IC News today.
By 
