Masked Symbol Modeling for Demodulation of Oversampled Baseband Communication Signals in Impulsive Noise-Dominated Channels

TL;DR

This paper introduces Masked Symbol Modeling (MSM), a Transformer-based framework that leverages the inherent context in oversampled baseband signals to improve demodulation in impulsive noise channels, moving towards more interpretative PHY layer processing.

Contribution

The paper presents MSM, a novel Transformer-based approach that uses masked-symbol prediction to learn the contextual structure of physical waveforms for robust demodulation.

Findings

MSM effectively predicts missing symbols in impulsive noise environments.

The approach captures the latent syntax of baseband waveforms.

Results indicate improved demodulation performance in challenging noise conditions.

Abstract

Recent breakthroughs in natural language processing show that attention mechanism in Transformer networks, trained via masked-token prediction, enables models to capture the semantic context of the tokens and internalize the grammar of language. While the application of Transformers to communication systems is a burgeoning field, the notion of context within physical waveforms remains under-explored. This paper addresses that gap by re-examining inter-symbol contribution (ISC) caused by pulse-shaping overlap. Rather than treating ISC as a nuisance, we view it as a deterministic source of contextual information embedded in oversampled complex baseband signals. We propose Masked Symbol Modeling (MSM), a framework for the physical (PHY) layer inspired by Bidirectional Encoder Representations from Transformers methodology. In MSM, a subset of symbol aligned samples is randomly masked, and a Transformer predicts the missing symbol identifiers using the surrounding "in-between" samples. Through this objective, the model learns the latent syntax of complex baseband waveforms. We illustrate MSM's potential by applying it to the task of demodulating signals corrupted by impulsive noise, where the model infers corrupted segments by leveraging the learned context. Our results suggest a path toward receivers that interpret, rather than merely detect communication signals, opening new avenues for context-aware PHY layer design.