Online Learning of Modular Bayesian Deep Receivers: Single-Step Adaptation with Streaming Data

2025-11-08

AI Summary

To address the poor adaptability of deep neural network (DNN) receivers under rapidly time-varying wireless channels, as well as the high latency and computational overhead of online stochastic gradient descent (SGD), this paper proposes a one-step adaptive online learning framework based on Bayesian parameter tracking. The method models online learning as Bayesian dynamic tracking in parameter space and integrates a modular DNN architecture to enable parallel, localized variational Bayesian updates—eliminating reliance on iterative gradient descent. This enables low-complexity, real-time channel estimation and signal detection on streaming communication data. Experiments demonstrate significantly reduced update latency, consistently low bit error rates, and enhanced robustness to abrupt channel variations. Overall, the proposed approach outperforms conventional SGD-based online learning schemes across all key metrics.

Technology Category

• Machine Learning: Bayesian Learning
• Natural Language Processing: Learning & Optimization for NLP
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Deep neural network (DNN)-based receivers offer a powerful alternative to classical model-based designs for wireless communication, especially in complex and nonlinear propagation environments. However, their adoption is challenged by the rapid variability of wireless channels, which makes pre-trained static DNN-based receivers ineffective, and by the latency and computational burden of online stochastic gradient descent (SGD)-based learning. In this work, we propose an online learning framework that enables rapid low-complexity adaptation of DNN-based receivers. Our approach is based on two main tenets. First, we cast online learning as Bayesian tracking in parameter space, enabling a single-step adaptation, which deviates from multi-epoch SGD . Second, we focus on modular DNN architectures that enable parallel, online, and localized variational Bayesian updates. Simulations with practical communication channels demonstrate that our proposed online learning framework can maintain a low error rate with markedly reduced update latency and increased robustness to channel dynamics as compared to traditional gradient descent based method.

Problem

• Enabling rapid adaptation of DNN-based receivers for wireless communications
• Overcoming computational burden of online SGD learning in dynamic channels
• Maintaining low error rates with reduced latency in varying environments

Innovation

• Online Bayesian tracking for single-step adaptation
• Modular DNN architectures enabling parallel updates
• Variational Bayesian updates reducing latency and complexity