Saved in:
Bibliographic Details
Main Authors: Rácz, András, Borsos, Tamás, Veres, András, Csala, Benedek
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.20363
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909865130065920
author Rácz, András
Borsos, Tamás
Veres, András
Csala, Benedek
author_facet Rácz, András
Borsos, Tamás
Veres, András
Csala, Benedek
contents We present AttDet, a Transformer-inspired MIMO (Multiple Input Multiple Output) detection method that treats each transmit layer as a token and learns inter-stream interference via a lightweight self-attention mechanism. Queries and keys are derived directly from the estimated channel matrix, so attention scores quantify channel correlation. Values are initialized by matched-filter outputs and iteratively refined. The AttDet design combines model-based interpretability with data-driven flexibility. We demonstrate through link-level simulations under realistic 5G channel models and high-order, mixed QAM modulation and coding schemes, that AttDet can approach near-optimal BER/BLER (Bit Error Rate/Block Error Rate) performance while maintaining predictable, polynomial complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2510_20363
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Transformer Inspired AI-based MIMO receiver
Rácz, András
Borsos, Tamás
Veres, András
Csala, Benedek
Signal Processing
Machine Learning
We present AttDet, a Transformer-inspired MIMO (Multiple Input Multiple Output) detection method that treats each transmit layer as a token and learns inter-stream interference via a lightweight self-attention mechanism. Queries and keys are derived directly from the estimated channel matrix, so attention scores quantify channel correlation. Values are initialized by matched-filter outputs and iteratively refined. The AttDet design combines model-based interpretability with data-driven flexibility. We demonstrate through link-level simulations under realistic 5G channel models and high-order, mixed QAM modulation and coding schemes, that AttDet can approach near-optimal BER/BLER (Bit Error Rate/Block Error Rate) performance while maintaining predictable, polynomial complexity.
title A Transformer Inspired AI-based MIMO receiver
topic Signal Processing
Machine Learning
url https://arxiv.org/abs/2510.20363