Defense Advanced Research Projects AgencyTagged Content List

Algorithms

A process or rule set used for calculations or other problem-solving operations

Showing 96 results for Algorithms RSS
Researchers have demonstrated effective attacks on machine learning (ML) algorithms. These attacks can cause high-confidence misclassifications of input data, even if the attacker lacks detailed knowledge of the ML classifier algorithm and/or training data. Developing effective defenses against such attacks is essential if ML is to be used for defense, security, or health and safety applications.
From phony news on Web sites to terrorist propaganda on social media to recruitment videos posted by extremists, conflict in the information domain is becoming a ubiquitous addition to traditional battlespaces. Given the pace of growth in social media and other networked communications, this bustling domain of words and images—once relegated to the sidelines of strategic planning—is poised to become ever more critical to national security and military success around the globe.
Most camera designers seek to maximize spatial resolution and signal-to-noise (SNR). A wealth of information in the optical domain, however, is lost under those constraints. Specialty cameras exist to capture other types of information, but are not normally able to provide high SNR imagery at high spatial resolution from a single focal plane, and are used infrequently due to demands of additional camera systems. Today’s imaging systems primarily perform a single or limited set of measurements due, in part, to the underlying readout integrated circuits (ROICs), which sample the signal of interest and transfer the values off of the chip. Typically, ROICs are designed for a specific mode of operation, and, in essence, are application specific integrated circuits (ASICs).
Current artificial intelligence (AI) systems excel at tasks defined by rigid rules – such as mastering the board games Go and chess with proficiency surpassing world-class human players. However, AI systems aren’t very good at adapting to constantly changing conditions commonly faced by troops in the real world – from reacting to an adversary’s surprise actions, to fluctuating weather, to operating in unfamiliar terrain.
Serial Interactions in Imperfect Information Games Applied to Complex Military Decision Making (SI3-CMD) builds on recent developments in artificial intelligence and game theory to enable more effective decisions in adversarial domains. SI3-CMD will explore several military decision making applications at strategic, tactical, and operational levels and develop AI/game theory techniques appropriate for their problem characteristics.