Rapport (Rapport Contrat/Projet) Année : 2022

AI4CODE Technical Report - Deliverable D3.1

Résumé

This deliverable reports on the activities carried out within the Task 3.1 of Work Package 3, aimed at establishing a design space exploration for ML-augmented decoding. Accordingly, the deliverable first formalizes several important communication scenarios where the performance or complexity of FEC decoders do not meet the expected requirements and could be improved by learning algorithms, and specifies the targeted improvements and corresponding key performance indicators. It further provides a systematic state of the art review of ML-based FEC decoders, summarizing recent advances in ML-augmented decoding of LDPC, Turbo, Polar, and more general linear block codes. Finally, it establishes a design space exploration for ML-augmented decoding, identifying parameters or technical components that can benefit from ML, and providing guidelines for improved designs of FEC decoders to be explored in the project.

Fichier principal
Vignette du fichier
AI4CODE_WP3_Livrable_D3_1_final.pdf (874.34 Ko) Télécharger le fichier

Dates et versions

hal-05333784 , version 1 (27-10-2025)

Licence

Identifiants

  • HAL Id : hal-05333784 , version 1

Citer

Raphaël Le Bidan, Charbel Abdel Nour, Elsa Dupraz, Catherine Douillard, Valentin Savin, et al.. AI4CODE Technical Report - Deliverable D3.1. IMT ATLANTIQUE. 2022. ⟨hal-05333784⟩
77 Consultations
154 Téléchargements

Partager

  • More