CrossMPT: Cross-attention Message-Passing Transformer for Error Correcting Codes

arXiv:2405.01033v1 Announce Type: new
Abstract: Error correcting codes~(ECCs) are indispensable for reliable transmission in communication systems. The recent advancements in deep learning have catalyzed the exploration of ECC decoders based on neural networks. Among these, transformer-based neural decoders have achieved state-of-the-art decoding performance. In this paper, we propose a novel Cross-attention Message-Passing Transformer~(CrossMPT). CrossMPT iteratively updates two types of input vectors (i.e., magnitude and syndrome vectors) using two masked cross-attention blocks. The mask matrices in these cross-attention blocks are determined …

abstract art arxiv attention communication cs.it cs.lg decoding deep learning error exploration math.it networks neural networks novel paper performance state systems transformer type