CSAFA labs’ research focused on the intersection of hardware-assisted cybersecurity, cyberattack-resilient technologies and foundational principles of machine learning. We generally prefer a more fundamental problem within a particular topic. Solving fundamental problem often resulted in new or enhanced capability that generalize beyond the specific application the research is focusing. Many of my research topics begin with analyzing and studying the problem at hand and identifying the fundamental/root-cause issues. Once the underlying issues are cleared, promising approaches and innovative solutions naturally emerge.
- What we are currently working on:
- Cyber attack resilient systems: Subprocess BFT++
- Heterogeneous lock-step processors: Hecocefta
- Real-time instruction level monitoring: MonT & Pokemon
- Software Distillation
- Hybrid reinforcement learning & transfer learning: MEAC
- LLM – Good & Evil
- Programming language translations: C2Rust
- Reasoning for exploits-chain: METALLIC