SPARK Neuro Familiarity Detector

The SPARK Neuro Inc. Familiarity Detector significantly enhances military capabilities in assessing the likelihood that a subject recognizes or has interacted with a target individual. Recent improvements in U.S. Signal Intelligence (SIGINT) capabilities have caused adversaries to revert to traditional tradecraft to plan their operations, including the widespread and growing use of couriers. As enemy Tactics, Techniques, and Procedures (TTPs) in use are increasingly analog in nature, threat networks that were once dissected through digital forensics have become obfuscated. This reduced digital footprint creates challenges both in (i) gathering intelligence on enemies to enable our global pursuit of dangerous adversaries and (ii) uncovering the intentions and plans of the individuals who make up our indigenous partner forces. Currently, there are limited market solutions to combat this increasing concern, and the current vetting process of indigenous partner forces is unreliable. DoD personnel in conflict zones and at home are in danger of potential insider threats, as demonstrated by the recent tragedy in the Pensacola Naval Air Station shooting conducted by a student from Saudi Arabia. In addition, there is a high risk of program failure given that the current process is highly subjective and relies on the skill of a trained interrogator, who may not be available when needed in theater or on a particular objective. The proposed project directly addresses the needs of various intelligence organizations and Special Operation Forces (SOF) within the Department of Defense and advances the goals of force protection, intelligence, and counterintelligence by leveraging AI to counter adversary tradecraft.  SPARK Neuro’s neuro-driven Familiarity Detector is addressing this problem by assessing the likelihood that a subject is familiar or has interacted with a target individual. The lightweight tactical kit includes an electroencephalogram (EEG) and a laptop pre-installed with our automatic processing software. The system rapidly presents images of faces to a human subject and automatically processes their neural signals for an immediate, concise visual presentation of the subject’s familiarity with every target shown. SPARK Neuro proposes to enhance its current Familiarity Detector through (i) improving the UI/UX of the interface, (ii) adding machine learning and computer vision to pre-process stimuli, and (iii) validating the system’s ability to perform effectively in operational environments.