可学习学习速率的无梯度联合梯度提升树（英文版）

英文标题：Gradient-less Federated Gradient Boosting Trees with Learnable Learning Rates中文摘要：提出一种水平联邦 XGBoost 的创新框架，通过使聚合树模型的学习率可学习，从而不依赖于梯度共享，同时提高隐私和通信效率，降低通信轮数和开销，并在各种数据集上进行了广泛评估。英文摘要：The privacy-sensitive nature of decentralized datasets and the robustness ofeXtreme Gradient Boosting (XGBoost) on tabular data raise the needs to trainXGBoost in the context of federated learning (FL). Existing works on federatedXGBoost in the horizontal setting rely on the sharing of gradients, whichinduce per-node