no code implementations • 28 Jun 2022 • Naifu Zhang, Meixia Tao, Jia Wang, Fan Xu
One of the main focuses in distributed learning is communication efficiency, since model aggregation at each round of training can consist of millions to billions of parameters.
no code implementations • 21 Jan 2021 • Naifu Zhang, Meixia Tao, Jia Wang
In FL, however, the model update is an indirect multi-terminal source coding problem, also called as the CEO problem where each edge device cannot observe directly the gradient that is to be reconstructed at the decoder, but is rather provided only with a noisy version.
no code implementations • 12 Sep 2020 • Nicholas Capel, Naifu Zhang
This paper proposes a hybrid reinforcement learning controller which dynamically interpolates a model-based linear controller and an arbitrary differentiable policy.
1 code implementation • 10 Mar 2020 • Yuhong Deng, Di Guo, Xiaofeng Guo, Naifu Zhang, Huaping Liu, Fuchun Sun
In this paper, we propose a novel task, Manipulation Question Answering (MQA), where the robot performs manipulation actions to change the environment in order to answer a given question.
no code implementations • 4 Mar 2020 • Naifu Zhang, Meixia Tao
We obtain the optimal policy in closed form when gradient statistics are given.