TY - RPRT TI - A Historical Trajectory Assisted Optimization Method for Zeroth-Order Federated Learning AU - Chenlin Wu AU - Xiaoyu He AU - Zike Li AU - Jing Gong AU - Zibin Zheng PY - 2024 UR - https://arxiv.org/abs/2409.15955 ID - 2409.15955 ER -