Synthetic and High-Fidelity Generation and Universal AI Use of Reliable Data (SHIFT-GUARD)

Future mission-critical operations will rely less on human and more on automated analysis using interconnected technologies that include a range of custom-designed, special purpose sensors installed on unmanned aerial vehicles (UAVs), the Internet of Battlefield Things (IoBT) sensors and cognitive radios for threat detection. Multi-domain data such as images and RF signals will be captured by each device and either processed at the edge or sent to a data fusion center to analyze the presence of potential threats. Artificial Intelligence (AI) / Machine learning (ML) has recently been broadly used to achieve surveillance tasks by analyzing either the recorded images captured by satellites or UAVs or over-the-air signals for anomaly detection. Since AI/ML algorithms require large amounts of data for training and testing, the availability of high-quality multi-domain datasets is vital to the mission success. There are many publicly available image and RF signal datasets that can be used to train and test ML algorithms for commercial tasks. However, representative datasets for mission-critical operations, such as images of drones, ships or tanks or wireless signals in different domains are very limited. Nexcepta Inc. proposes to develop the Synthetic & High-Fidelity Generation and Universal AI Use of Reliable Data (SHIFT-GUARD) technology which focuses on the synthetic data generation (both data augmentation and domain adaptation) using AI-based methods. SHIFT-GUARD will be a game changer for generating multi-domain high-quality synthetic data and a catalyzer for the development of high accuracy ML algorithms which will both benefit both mission-critical operations and commercial field.