Learning to Detect
Learning to Detect
复制标题
DOI：
10.1109/tsp.2019.2899805
复制
发表时间：
2018-05
影响因子：
5.4
通讯作者：
N. Samuel;Tzvi Diskin;A. Wiesel
中科院分区：
文献类型：
--
作者：
N. Samuel;Tzvi Diskin;A. Wiesel
在本文中，我们考虑使用深度神经网络进行多输入多输出检测。我们介绍了两种不同的深度架构：标准的全连接多层网络和专门为该任务设计的检测网络（DetNet）。DetNet的结构是通过将投影梯度下降算法的迭代展开到网络中来获得的。我们比较了所提出的方法的准确性和运行时的复杂性，并在保持低计算要求的同时实现了最先进的性能。此外，我们设法训练单个网络来检测整个渠道分布。最后，我们考虑软输出的检测，并表明网络可以很容易地进行修改，以产生软决策。
In this paper, we consider multiple-input-multiple-output detection using deep neural networks. We introduce two different deep architectures: a standard fully connected multi-layer network, and a detection network (DetNet), which is specifically designed for the task. The structure of DetNet is obtained by unfolding the iterations of a projected gradient descent algorithm into a network. We compare the accuracy and runtime complexity of the proposed approaches and achieve state-of-the-art performance while maintaining low computational requirements. Furthermore, we manage to train a single network to detect over an entire distribution of channels. Finally, we consider detection with soft outputs and show that the networks can easily be modified to produce soft decisions.