Structured Masked Diffusion for Joint Multiuser Decoding

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lee, Taekyun, Yun, Jiyoung, Andrews, Jeffrey G., Kim, Hyeji
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916047115780096
author Lee, Taekyun
Yun, Jiyoung
Andrews, Jeffrey G.
Kim, Hyeji
author_facet Lee, Taekyun
Yun, Jiyoung
Andrews, Jeffrey G.
Kim, Hyeji
contents In joint multiuser decoding, a receiver recovers a set of messages from a single noisy aggregate of many simultaneous transmissions. Classical decoders rely on rule-based mechanisms such as successive interference cancellation, joint belief propagation, or list recovery, all of which become brittle or expensive as ambiguity increases. We propose CIDER, a learned multiuser decoder with masked-diffusion refinement steps. CIDER uses demixing to prevent duplicate-row collapse and uses parity-aware propagation to provide soft guidance from the code constraints. In higher-load regimes, we further improve reliability via a lightweight quality-guided remasking step that selectively re-decodes low-confidence sequences. On commonly used error-correcting codes, CIDER matches or improves on FFT-accelerated joint belief propagation-style decoding in symbol error rate while running more than $6\times$ to over $100\times$ faster, with the speedup widening as the blocklength grows. Code is available at https://github.com/jiyunyoung/CIDER.
format Preprint
id arxiv_https___arxiv_org_abs_2605_26580
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Structured Masked Diffusion for Joint Multiuser Decoding
Lee, Taekyun
Yun, Jiyoung
Andrews, Jeffrey G.
Kim, Hyeji
Information Theory
Signal Processing
In joint multiuser decoding, a receiver recovers a set of messages from a single noisy aggregate of many simultaneous transmissions. Classical decoders rely on rule-based mechanisms such as successive interference cancellation, joint belief propagation, or list recovery, all of which become brittle or expensive as ambiguity increases. We propose CIDER, a learned multiuser decoder with masked-diffusion refinement steps. CIDER uses demixing to prevent duplicate-row collapse and uses parity-aware propagation to provide soft guidance from the code constraints. In higher-load regimes, we further improve reliability via a lightweight quality-guided remasking step that selectively re-decodes low-confidence sequences. On commonly used error-correcting codes, CIDER matches or improves on FFT-accelerated joint belief propagation-style decoding in symbol error rate while running more than $6\times$ to over $100\times$ faster, with the speedup widening as the blocklength grows. Code is available at https://github.com/jiyunyoung/CIDER.
title Structured Masked Diffusion for Joint Multiuser Decoding
topic Information Theory
Signal Processing
url https://arxiv.org/abs/2605.26580