The Lancelot framework was jointly proposed in September 2025 by a research team from the AIoT Lab of the Chinese University of Hong Kong, Chongqing University, City University of Hong Kong, and other universities and institutions. The relevant research results were published in the paper "Towards compute-efficient Byzantine-robust federated learning with fully homomorphic encryption".

Lancelot is a computationally efficient framework for privacy-preserving Byzantine Robust Federated Learning (BRFL) under fully homomorphic encryption (FHE). Specifically, researchers propose a novel interactive federated learning paradigm for decentralized collaboration among institutions, enabling the training of high-performance and robust models without sacrificing data privacy and computational efficiency. Lancelot combines algorithmic enhancements and hardware acceleration, covering optimizations for pairwise ciphertext multiplication policies, polynomial matrix multiplication, and complex addition operations.