In response to the Navys solicitation for cutting-edge artificial intelligence (AI) and machine learning (ML) technologies from innovative small businesses to address counter AI, we propose a novel and trustworthy AI solution named TrustAI. TrustAI has a new model structure and corresponding fast learning algorithms to design a recognizer or classifier based on application requirements. The trained TrustAI structure is transparent and explainable. It can be used to replace DNN, CNN, and other existing neural-network-based machine learning approaches for all military and commercial applications. Furthermore, we propose a new validation and verification method in order to evaluate the performance of AI systems in terms of trustworthy and robustness. TrustAI has the capability to resolve most known counter AI attacks as reported in the solicitation and the literature.
Benefit: The proposed trustworthy AI system has broad applications in both military and commercial markets. For military applications, we can license the TrustAI software to prime contractors and work with them to develop real-world applications and products. For the commercial market, the potential market is huge. For example, automobile manufacturers can use TrustAI to improve self-driving cars; aircraft manufacturers and airlines can use TrustAI for autopilot systems; the medical field can use TrustAI for healthcare decisions; financial and insurance companies can use TrustAI for credit and investment decisions, etc. According to Tractica, the global artificial intelligence software market is expected to experience massive growth in the coming years, with revenues increasing from around $9.5 billion U.S. dollars in 2018 to an expected $118.6 billion by 2025. The TrustAI software can be used as the core for all above AI applications. LCT has a successful track record in the commercialization of SBIR technology.
Keywords: Robustness, Robustness, counter AI, Machine Learning, decision, Artificial Intelligence, Neural networks, classifier, trustworthy AI
