Research
My research examines the security and privacy of federated, decentralized, and edge AI, particularly when devices or participants cannot be fully trusted.
Verifiable and post-quantum distributed learning
I develop zero-knowledge methods for verifying evaluation thresholds without revealing local metrics, and post-quantum authentication mechanisms for distributed learning. I also study federated graph learning for cross-institutional analysis without sharing raw transaction graphs.
Representative work: ZKP-FedEval, PQS-BFL, and FedGraph-VASP.
Resilience against malicious and strategic participants
I study incentive mechanisms intended to discourage poisoning by strategic participants, and adaptive honeypot deployment for resource-constrained IoT systems.
Representative work: Bayesian incentive mechanisms and blockchain-enabled honeypot conversion.
Cryptographic assurance for emerging systems
I combine physical unclonable functions (PUFs) with zero-knowledge proofs to authenticate devices without exposing their underlying identity secrets. I also evaluate the performance and deployment tradeoffs of post-quantum cryptography on consumer and edge hardware.
Representative work: PUFZIN and post-quantum cryptography for consumer electronics.