sunyaping
  • Joined on Oct 22, 2021
  • Organization
Loading Heatmap…

sunyaping pushed to master at AINET/Networks-For-AI

  • 753f630bf4 Meta-Reinforcement Learning for Trajectory Design in Wireless UAV Networks This is the code for the paper ''Meta-Reinforcement Learning for Trajectory Design in Wireless UAV Networks''. Please run this simulation using python=3.6.5 or above. Please install the gym package before the implementation.

7 months ago

sunyaping pushed to master at AINET/Networks-For-AI

  • 1f3f8698ba Convergence Time Minimization of Federated Learning over Wireless Networks This is the code for the paper ''Convergence Time Minimization of Federated Learning over Wireless Networks''. Please run this simulation using Matlab 2018b or above. Please install the Machine Learning toolbox at Matlab before the implementation. To run this simulation, please first download the MNIST dataset from http://yann.lecun.com/exdb/mnist/.

7 months ago

sunyaping pushed to master at AINET/Networks-For-AI

7 months ago

sunyaping pushed to master at AINET/Networks-For-AI

  • 6846e47cdf Joint User Association and Resource Allocation Optimization for MEC-Enabled IoT Networks Technical report for "Joint User Association and Resource Allocation Optimization for MEC-Enabled IoT Networks".

7 months ago

sunyaping pushed to master at AINET/AI-For-Networks

  • 7b41fbfb8a 分布式高斯过程模型 该分布式高斯过程模型包含两个算法,即基于矩阵近似的分布式GP训练算法和基于交叉验证的分布式GP预测算法,分别对应于高斯过程的训练过程和预测过程。成果发表于IEEE Journal on Selected Areas in Communications.上的期刊论文Wireless Traffic Prediction with Scalable Gaussian Process: Framework, Algorithms, and Verification。分布式GP训练算法的核心思想是基于ADMM算法将原高斯过程的超参数优化问题拆分为多个更易于解决的子问题,并交由多个分布式计算单元分别解决。分布式GP预测算法的核心思想是利用交叉验证思想,将分布式计算单元的本地预测结果融合为一个更准确的全局预测结果。

8 months ago

sunyaping pushed to master at AINET/AI-For-Networks

  • 23fd278b09 基于语义通信的联合编码项目:传统方法Baseline部分代码 基于语义通信的联合编码项目:传统方法Baseline部分代码

8 months ago

sunyaping pushed to master at AINET/AI-For-Networks

  • 49bd8d899b 基于强化学习的自适应路由算法 基于强化学习的自适应路由算法

8 months ago

sunyaping pushed to master at AINET/AI-For-Networks

8 months ago

sunyaping pushed to master at AINET/AI-For-Networks

8 months ago

sunyaping pushed to master at AINET/Networks-For-AI

8 months ago

sunyaping pushed to master at AINET/Networks-For-AI

  • bc00858aeb IEEE TWC paper 实现代码 IEEE TWC paper "One-Bit Over-the-Air Aggregation for Communication-Efficient Federated Edge Learning- Design and Convergence Analysis" 的实现代码

8 months ago

sunyaping pushed to master at AINET/Networks-For-AI

  • a90ef44ab6 IEEE TWC paper IEEE TWC paper "One-Bit Over-the-Air Aggregation for Communication-Efficient Federated Edge Learning- Design and Convergence Analysis" 论文原文

8 months ago

sunyaping pushed to master at AINET/AI-and-Machine-Learning-Core-A...

  • 2da1efafe8 上传文件至 'lyqun' 尽管遮挡在自然界广泛存在,并且仍然是姿势估计的基本挑战,但是现有的基于热图的方法在遮挡上会严重退化。他们的内在问题是他们根据视觉信息直接定位关节。但是,看不见的关节却缺乏这种能力。与本地化相反,我们的框架通过提出一个图像指导的渐进式GCN模块从推理角度估计了不可见的关节,该模块提供了对图像上下文和姿势结构的全面理解。而且,现有的基准测试包含有限的评估。因此,我们彻底解决了这个问题,并提出了一个新颖的OPEC-Net框架以及一个带有9k批注图像的新的遮挡姿势(OCPose)数据集。对基准的大量定量和定性评估表明,OPEC-Net与最近的领先成果相比有了显着改进。值得注意的是,就相邻实例之间的平均IoU而言,我们的OCPose是最复杂的遮挡数据集。源代码和OCPose将公开可用

8 months ago

sunyaping pushed to master at AINET/AI-and-Machine-Learning-Core-A...

  • 4c4b20ab71 上传文件至 'lyqun' FPConv,这是一种专为3D点云分析而设计的新颖表面样式卷积运算符。 与以前的方法不同,FPConv不需要转换为中间表示形式(例如3D网格或图形),而是直接在点云的表面几何上工作。 更具体地说,对于每个点,FPConv通过自动学习权重图以将周围的点柔和地投影到2D网格上来执行局部展平。 因此可以将规则的2D卷积应用于有效的特征学习。 FPConv可以轻松集成到各种网络体系结构中,以执行3D对象分类和3D场景分割等任务,并可以与现有的体积类型卷积实现可比的性能。 更重要的是,我们的实验还表明FPConv可以作为体积卷积的补充,联合训练它们可以进一步提高整体性能,从而达到最新的结果

8 months ago

sunyaping pushed to master at AINET/AI-and-Machine-Learning-Core-A...

8 months ago

sunyaping pushed to master at AINET/Networks-For-AI

  • 25030d270d 上传文件至 'One-bit-over-the-air-computation-master' "One-Bit Over-the-Air Aggregation for Communication-Efficient Federated Edge Learning- Design and Convergence Analysis" 的实现代码

8 months ago

sunyaping pushed to master at AINET/Networks-For-AI

  • 22b65028d4 上传文件至 'One-Bit Over-the-Air Aggregation for Communication-Efficient Federated Edge Learning- Design and Convergence Analysis' IEEE TWC paper "One-Bit Over-the-Air Aggregation for Communication-Efficient Federated Edge Learning- Design and Convergence Analysis" 论文原文

8 months ago

sunyaping pushed to master at AINET/Networks-For-AI

8 months ago