CySense: Hardware-assisted Cyberattack Detection Engine

Cyberattacks are a growing concern across industry and government, with threats like ransomware, spyware, and more causing intellectual property theft, system downtime, and even fatalities. Most cyberattack detection techniques are limited in scalability and detection efficacy as the cyberattack landscape is constantly widening and evolving. Derivative variants of pre-existing cyberattacks are easily created that bypass existing detection techniques, and the protections have limited visibility into the full cyber system leading to high false-positive rates. Caspia's proposed work, CySense, will develop a dynamic and advanced cyberattack detection tool that harnesses the interaction of hardware-assisted technology and state-of-the-art machine learning to pinpoint and neutralize malicious or irregular computing system behavior. CySense innovates by harnessing low-level performance data generated by modern hardware and computing components, thereby uncovering subtle traces of malicious intent. This innovative methodology empowers CySense to proactively identify and interdict cyberattacks in real-time, ensuring the protection of users against a spectrum of both well-known and emerging threats. In Phase I, Caspia successfully demonstrated the feasibility of machine learning techniques and hardware-generated data to accurately detect anomalous software behavior from real cyberattacks executed on Caspia's real test platform. In Phase II, Caspia will demonstrate the prototype CySense solution in a real-world or simulated test environment with the proposed detection pipeline and enhanced machine learning algorithms. A roadmap that takes the program through Phase III will be part of the Phase II delivery. The proposed activities will progress commercialization goals of offering a highly accurate and real-time detection engine for critical cyber-threats.