The challenging neural decoding with general-purpose networks and its improvement via probabilistic embeddings - Télécom Paris Accéder directement au contenu
Poster De Conférence Année : 2023

The challenging neural decoding with general-purpose networks and its improvement via probabilistic embeddings

Résumé

As deep learning is becoming more and more popular in a variety of fields, it is natural to ask whether it can be used for decoding error-correcting codes. In this work, we show that the answer is yes, but with a caveat: Naive application of general-purpose neural networks is not well suited for bitwise decoding. We thoroughly identify the challenges that prevent general-purpose neural networks from decoding successfully, including the curse of dimensionality and the requirement of extremely high accuracy. We then propose a proba- bilistic embedding method in the preprocessing stage that facilitate the learning. We also show that this new method allows general-purpose neural networks to decode with a perfor- mance that is close to the theoretical optimality while saving time compared to traditional decoding methods such as maximum-likelihood decoding or the BCJR algorithm.
Fichier non déposé

Dates et versions

hal-04416031 , version 1 (25-01-2024)

Identifiants

  • HAL Id : hal-04416031 , version 1

Citer

Xiaolin Wang, Joseph J. Boutros, Olivier Rioul. The challenging neural decoding with general-purpose networks and its improvement via probabilistic embeddings. 8th Junior Conference on Data Science and Engineering (JDSE'23), Sep 2023, Orsay (Université Paris-Sud 11), France. ⟨hal-04416031⟩
16 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More