Multi-layer MIMO Relay as Deep Physical Neural Networks: Power Amplifiers as Activation Functions

AI 中文解读

**核心亮点**：科学家用无线通信中的“信号中继站”搭建成一个物理神经网络，巧妙地利用功率放大器的天然特性来完成AI计算，省电又快速。 **通俗解读**：传统AI依赖数字芯片进行海量计算，功耗大且延时高。这项研究另辟蹊径——它让多个WIFI信号中继器串联起来，每个中继器在放大信号时，自身硬件固有的“变形”（即非线性失真）恰恰充当了神经网络中的“激活函数”。信号在中继间层层传递，就像数据在神经网络中逐层处理一样。而且这些中继器的参数可以通过算法自动优化，整个系统无需专门AI芯片就能完成推理任务，比如识别图片内容。 **实际影响**：这意味着未来的基站、路由器甚至卫星，有可能在传输信号的同时“顺便”完成AI计算。对普通人来说，最直接的好处是设备更省电、响应更迅速——比如智能摄像头无需上传云端就能本地识别动作，物联网传感器也能实时分析数据而不耗电过多。长远看，这项技术有望推动“空中智能网络”的普及，让通信基础设施本身具备AI能力，降低对高性能芯片的依赖。

arXiv:2607.18354v1 Announce Type: new Abstract: Wireless physical neural networks (WPNNs) embed neural computation directly into analog hardware, offering lower energy consumption and latency than conventional digital implementations. In this paper, we propose a deep WPNN in which nonlinear activations are realized by a multi-hop multiple-input multiple-output (MIMO) relay network, in which each relay implements a trainable complex linear gain and bias, followed by the power amplifier's intrinsic nonlinearity acting as an activation function. The cascade of multiple relays therefore realizes an over-the-air fully connected network whose parameters can be trained end-to-end. We develop two transceiver designs for different channel state information (CSI) availability scenarios: a least squares (LS)-based scheme requiring only receiver-side CSI, and a singular-value-decomposition (SVD)-based scheme requiring both transmitter-side and receiver-side CSI. Simulation results show that the proposed architecture enables accurate over-the-air inference for image classification. In particular, the results highlight the advantage of exploiting hardware nonlinearity for enhanced inference capability.