With the boom of technologies such as IoT and AI, security is increasingly becoming a requirement rather than an option.  Given the completely different attack surfaces and the physical constraints of smart IoT devices, the direct adoption of conventional cyber-security solutions is misleading. Lightweight security solutions that provide the root of trust from hardware and comprehensive protections from data sources, transmission channels, processing, and user interactions are urged. We are interested in investigating the security and privacy vulnerabilities in smart IoT systems and developing corresponding countermeasures. In particular, we are interested at 

1. Hardware security primitives, including physical unclonable functions, true random number generators, radio frequency fingerprint, etc. 

2. Lightweight end-to-end security design for smart systems, including mutual authentication and key exchange protocols, group key protocols, etc. 

3. Neural network security and privacy: attacks and defenses, including adversarial attacks, backdoor attacks, model IP protections, etc.