Grad-FEC: Unequal Loss Protection of Deep Features in Collaborative Intelligence

TL;DR

This paper introduces Grad-FEC, a method that enhances collaborative intelligence systems by selectively protecting important deep features with error correction to improve robustness against packet loss.

Contribution

It proposes a novel Unequal Loss Protection approach using feature importance estimation and selective FEC coding for resilient CI systems.

Findings

Significant improvement in system robustness under packet loss

Effective identification of critical feature packets for protection

Enhanced reliability of deep feature transmission in CI

Abstract

Collaborative intelligence (CI) involves dividing an artificial intelligence (AI) model into two parts: front-end, to be deployed on an edge device, and back-end, to be deployed in the cloud. The deep feature tensors produced by the front-end are transmitted to the cloud through a communication channel, which may be subject to packet loss. To address this issue, in this paper, we propose a novel approach to enhance the resilience of the CI system in the presence of packet loss through Unequal Loss Protection (ULP). The proposed ULP approach involves a feature importance estimator, which estimates the importance of feature packets produced by the front-end, and then selectively applies Forward Error Correction (FEC) codes to protect important packets. Experimental results demonstrate that the proposed approach can significantly improve the reliability and robustness of the CI system in the presence of packet loss.