Deep Joint Source-Channel Coding for Multi-Task Network

TL;DR

This paper introduces a multi-task learning network with deep joint source-channel coding for collaborative intelligence, achieving high compression and robustness over wireless channels while maintaining performance on object detection and segmentation.

Contribution

It proposes a novel feature fusion multi-task network split between mobile and edge, combined with deep JSCC for efficient, robust transmission in collaborative intelligence scenarios.

Findings

Achieves 512x compression of intermediate features.

Maintains within 2% performance loss on tasks.

Outperforms separate source and channel coding schemes.

Abstract

Multi-task learning (MTL) is an efficient way to improve the performance of related tasks by sharing knowledge. However, most existing MTL networks run on a single end and are not suitable for collaborative intelligence (CI) scenarios. In this work, we propose an MTL network with a deep joint source-channel coding (JSCC) framework, which allows operating under CI scenarios. We first propose a feature fusion based MTL network (FFMNet) for joint object detection and semantic segmentation. Compared with other MTL networks, FFMNet gets higher performance with fewer parameters. Then FFMNet is split into two parts, which run on a mobile device and an edge server respectively. The feature generated by the mobile device is transmitted through the wireless channel to the edge server. To reduce the transmission overhead of the intermediate feature, a deep JSCC network is designed. By combining two networks together, the whole model achieves 512x compression for the intermediate feature and a performance loss within 2% on both tasks. At last, by training with noise, the FFMNet with JSCC is robust to various channel conditions and outperforms the separate source and channel coding scheme.