Feng Guo

FedAttr introduces a novel client-level attribution protocol for Federated Learning (FL) that accurately identifies which clients trained on watermarked data while maintaining strong privacy guarantees.

The paper introduces SeedHijack, a backdoor attack that manipulates the pseudorandom number generation process in LLMs to force specific token selections, and proposes a hardware quantum random number generator (QRNG) as a robust defense.

This paper proposes DPIFrame, a dual parallelizable framework for accelerating Click-through rate (CTR) model inference on GPU, achieving state-of-the-art inference performance with significant speedups.