X-REFINE: XAI-based RElevance input-Filtering and archItecture fiNe-tuning for channel Estimation

Abdul Karim Gizzini; Yahia Medjahdi

cs.LGFeb 25, 2026

X-REFINE: XAI-based RElevance input-Filtering and archItecture fiNe-tuning for channel Estimation

Authors: Abdul Karim Gizzini, Yahia Medjahdi

Organizations: University of Paris-Est Créteil (UPEC), LISSI/TincNET, F-94400, Vitry-sur-Seine, France. · IMT Nord Europe, Institut Mines Télécom, Centre for Digital Systems, F-59653 Villeneuve d’Ascq, France.

Abstract

AI-native architectures are vital for 6G wireless communications. The black-box nature and high complexity of deep learning models employed in critical applications, such as channel estimation, limit their practical deployment. While perturbation-based eXplainable Artificial Intelligence (XAI) solutions offer input filtering, they often neglect internal structural optimization. We propose X-REFINE, an XAI-based framework for joint input-filtering and architecture fine-tuning. By utilizing a decomposition-based, sign-stabilized LRP epsilon rule, X-REFINE backpropagates predictions to derive high-resolution relevance scores for both subcarriers and hidden neurons. This enables a reliable optimization that identifies the most reliable model components. Simulation results demonstrate that X-REFINE achieves a superior performance-complexity-interpretability trade-off compared to the external perturbation-based XAI frameworks, significantly reducing computational complexity while maintaining robust bit error rate (BER) performance.

Figures & tables

Figure 1: Block diagram of the proposed X-REFINE framework.

Figure 2: Optimized performance of the proposed X-REFINE framework employing HF channel model and QPSK modulation order.

Figure 3: Relevance scores distribution of the proposed X-REFINE framework considering LF and HF channel models under different modulation orders: From left to right: (a) LF channel model, (b) HF channel model.

Figure 4: BER performance of the proposed X-REFINE framework for the LF and HF channel models under different modulation orders: From left to right: (a) LF - QPSK modulation, (b) LF - 64QAM modulation, (c) HF - QPSK modulation, and (d) HF - 64QAM modulation.

Figure 5: Relevance scores distribution of the benchmarked XAI schemes considering the HF channel model employing QPSK modulation order.

Figure 6: Performance evaluation of the benchmarked XAI methods for HF channel model employing QPSK modulation order: (a) BER, (b) Throughput.

Method	Complexity Reduction ( $\Delta\mathcal{C}$ )				$\mathcal{C}_{\text{XAI}}$
	LF		HF
	QPSK	64QAM	QPSK	64QAM
LIME	39.51%		18.11%	9.88%	$\mathcal{O}(D_{\text{LIME}}K_{\text{on}}^{2})$
DeepSHAP			8.23%	14.82%	$\mathcal{O}(D_{\text{SHAP}}K_{\text{on}})$
XAI-CHEST			24.70%	32.93%	$\mathcal{O}(ED_{\text{XAI-CHEST}}K_{\text{on}}^{2})$ + $\mathcal{O}(K_{\text{on}})$
X-REFINE	62.41%	51.41%	35.16%	43.59%	$\mathcal{O}(K_{\text{on}})$

Table I: Computational complexity comparison of

\mathcal{F}^{*}

optimized by different XAI methods.

Channel Estimation

Channel estimation infers wireless propagation effects from pilots or received signals so transmitters and receivers can communicate reliably.

See all 38 papers

X-RACE: XAI-assisted Recurrent neural network Attribution for Channel Estimation
Sep 10, 2026Abdul Karim Gizzini, Yahia MedjahdiEfficient Neural Network Inference Explainable Artificial Intelligence
University of Paris-Est Cr´eteil (UPEC), LISSI/TincNET, F-94400, Vitry-sur-Seine, France. · IMT Nord Europe, Institut Mines T´el´ecom, Centre for Digital Systems, F-59653 Villeneuve d’Ascq, France.
REACH: Interpretability-Driven Feature Identification and Architecture Compression for Multi-Channel Vehicular Channel Estimation
Jun 10, 2026Simbarashe Aldrin Ngorima, Albert Helberg, Marelie H. DavelFeature Attribution OOD Generalization
Faculty of Engineering, North-West University, South Africa · Centre for Artificial Intelligence Research, South Africa · National Institute for Theoretical and Computational Sciences, South Africa
Hardware-in-the-Loop Phase-Aware CNN for Real-Time 5G Channel Estimation
Aug 11, 2026Javad Zolfaghari-Bengar, Rakibul Rony, Elisa Gomez-de-Lope +3Wireless Communications Channel Estimation
Keysight AI Labs

Sep 10, 2026J/K move · Enter open · S save

X-RACE: XAI-assisted Recurrent neural network Attribution for Channel Estimation

Abdul Karim Gizzini, Yahia Medjahdi

University of Paris-Est Cr´eteil (UPEC), LISSI/TincNET, F-94400, Vitry-sur-Seine, France. · IMT Nord Europe, Institut Mines T´el´ecom, Centre for Digital Systems, F-59653 Villeneuve d’Ascq, France.

Efficient Neural Network Inference Explainable Artificial Intelligence Long Short-Term Memory Networks Channel Estimation Neural Network Pruning

Open paper AskPDF arXiv

Deep learning models, notably Long Short-Term Memory (LSTM), have demonstrated promising performance in channel estimation for high-mobility vehicular environments. However, their black-box nature and architectural overhead limit trustworthiness and efficiency. Classical explainable AI (XAI) methods rely on costly iterative processes, offering only input-level filtering without addressing architectural fine-tuning. To overcome these limitations, this paper proposes the XAI-assisted Recurrent neural network Attribution for Channel Estimation (X-RACE) framework. X-RACE uses a low-complexity, one-shot dual-optimization strategy to simultaneously evaluate and prune irrelevant input subcarriers and internal hidden units. Furthermore, we propose novel temporal XAI metrics: Saturation Time, Importance Drift, and Relevance Contrast to characterize the LSTM's learning dynamics and memory convergence. Extensive simulations demonstrate that X-RACE reduces inference complexity by at least 44.1% while improving or preserving Bit Error Rate (BER) performance, outperforming classical XAI schemes.

Abstract

Channel Estimation

Explore similar work

X-RACE: XAI-assisted Recurrent neural network Attribution for Channel Estimation

REACH: Interpretability-Driven Feature Identification and Architecture Compression for Multi-Channel Vehicular Channel Estimation

Hardware-in-the-Loop Phase-Aware CNN for Real-Time 5G Channel Estimation