Realization of a Fully Connected Neural Layer Over-the-Air through Multi-hop Amplify-and-Forward Relays

TL;DR

This paper demonstrates how to implement a fully-connected neural network layer over wireless channels using multi-hop amplify-and-forward relays, optimizing system parameters for high accuracy.

Contribution

It introduces an optimization framework for over-the-air neural network implementation with multi-hop relays, achieving near-perfect classification accuracy.

Findings

Multi-hop relaying achieves almost perfect classification accuracy.

Optimization of precoder, combiner, and relay gains enhances neural network performance.

Simulation results validate the effectiveness of the proposed over-the-air implementation.

Abstract

We study the problem of implementing a fully-connected layer of a neural network using wireless over-the-air computing. We assume a multi hop system with a multi-antenna transmitter and receiver, along with a number of multi-hop amplify-and-forward relay devices in between. We formulate an optimization problem that optimizes the transmitter precoder, receiver combiner and amplify-and-forward gains, subject to relay device power constraint and transmitter power constraint. We propose an alternating optimization framework that optimizes the imitation accuracy. Simulation study results reveal that multi-hop relaying achieves an almost perfect classification accuracy when used in a neural network.