# On Model Coding for Distributed Inference and Transmission in Mobile Edge Computing Systems

## TL;DR

This paper explores how coding model information in mobile edge computing can reduce latency in distributed linear inference tasks despite non-deterministic EN processing times.

## Contribution

It provides an information-theoretic analysis showing that coding model data can significantly lower total latency in distributed inference systems.

## Key findings

- Coding reduces overall latency in distributed inference.
- Cooperative transmission benefits are limited by coding.
- Coding is crucial for latency reduction despite non-deterministic EN times.

## Abstract

Consider a mobile edge computing system in which users wish to obtain the result of a linear inference operation on locally measured input data. Unlike the offloaded input data, the model weight matrix is distributed across wireless Edge Nodes (ENs). ENs have non-deterministic computing times, and they can transmit any shared computed output back to the users cooperatively. This letter investigates the potential advantages obtained by coding model information prior to ENs' storage. Through an information-theoretic analysis, it is concluded that, while generally limiting cooperation opportunities, coding is instrumental in reducing the overall computation-plus-communication latency.