Targeted Fine-Tuning of DNN-Based Receivers via Influence Functions

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Tuononen, Marko, Penttinen, Heikki, Hautamäki, Ville
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908778071326720
author Tuononen, Marko
Penttinen, Heikki
Hautamäki, Ville
author_facet Tuononen, Marko
Penttinen, Heikki
Hautamäki, Ville
contents We present the first use of influence functions for deep learning-based wireless receivers. Applied to DeepRx, a fully convolutional receiver, influence analysis reveals which training samples drive bit predictions, enabling targeted fine-tuning of poorly performing cases. We show that loss-relative influence with capacity-like binary cross-entropy loss and first-order updates on beneficial samples most consistently improves bit error rate toward genie-aided performance, outperforming random fine-tuning in single-target scenarios. Multi-target adaptation proved less effective, underscoring open challenges. Beyond experiments, we connect influence to self-influence corrections and propose a second-order, influence-aligned update strategy. Our results establish influence functions as both an interpretability tool and a basis for efficient receiver adaptation.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15950
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Targeted Fine-Tuning of DNN-Based Receivers via Influence Functions
Tuononen, Marko
Penttinen, Heikki
Hautamäki, Ville
Machine Learning
Signal Processing
68T05 (Primary), 94A05 (Secondary)
I.2.6; C.2.1; I.5.2
We present the first use of influence functions for deep learning-based wireless receivers. Applied to DeepRx, a fully convolutional receiver, influence analysis reveals which training samples drive bit predictions, enabling targeted fine-tuning of poorly performing cases. We show that loss-relative influence with capacity-like binary cross-entropy loss and first-order updates on beneficial samples most consistently improves bit error rate toward genie-aided performance, outperforming random fine-tuning in single-target scenarios. Multi-target adaptation proved less effective, underscoring open challenges. Beyond experiments, we connect influence to self-influence corrections and propose a second-order, influence-aligned update strategy. Our results establish influence functions as both an interpretability tool and a basis for efficient receiver adaptation.
title Targeted Fine-Tuning of DNN-Based Receivers via Influence Functions
topic Machine Learning
Signal Processing
68T05 (Primary), 94A05 (Secondary)
I.2.6; C.2.1; I.5.2
url https://arxiv.org/abs/2509.15950