Online Label Recovery for Deep Learning-based Communication through Error Correcting Codes

TL;DR

This paper introduces a method to use error correcting codes for online label recovery, enabling adaptive training of communication systems without dedicated known symbol transmission, thus improving robustness to channel variations.

Contribution

It proposes a novel approach to generate labels from ECCs for real-time training of communication systems, applicable to both end-to-end and partially trainable models.

Findings

Effective label recovery using ECCs for adaptive training.

Training with correct labels is crucial for system performance.

Extension of OFDM with trainable pre-equalizer neural network demonstrated.

Abstract

We demonstrate that error correcting codes (ECCs) can be used to construct a labeled data set for finetuning of "trainable" communication systems without sacrificing resources for the transmission of known symbols. This enables adaptive systems, which can be trained on-the-fly to compensate for slow fluctuations in channel conditions or varying hardware impairments. We examine the influence of corrupted training data and show that it is crucial to train based on correct labels. The proposed method can be applied to fully end-to-end trained communication systems (autoencoders) as well as systems with only some trainable components. This is exemplified by extending a conventional OFDM system with a trainable pre-equalizer neural network (NN) that can be optimized at run time.