Lstm channel estimation improved by xai driven pruning and metrics
X-RACE: XAI-assisted Recurrent neural network Attribution for Channel Estimation
Artificial Intelligence
Summary
Deep learning models like LSTMs help estimate communication channels in moving vehicles but are hard to understand and often inefficient. The authors created a new method called X-RACE that uses explainable AI to identify and remove unnecessary parts of the model and its inputs in one step, making it simpler and more trustworthy. They also introduced new ways to measure how the model learns over time. Their tests show X-RACE cuts computation by nearly half while keeping or improving data accuracy compared to older methods.
What this means in practice
- •For vehicular communication engineers: Optimize channel estimation models for high-mobility environments by reducing complexity and enhancing reliability using X-RACE pruning and analysis.
- •For wireless network designers: Improve real-time signal processing by applying low-complexity model pruning that preserves error rates in dynamic wireless channels.
Authors
Abdul Karim Gizzini, Yahia Medjahdi
Abstract
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.