Context-Aware Markov VAE for CSI Compression in Wireless Systems
2026-06-15 • Information Theory
Information TheoryMachine Learning
AI summaryⓘ
The authors study how to compress information about changing wireless signals, called channel state information (CSI), in systems with many antennas. They note that CSI changes over time in a predictable way, but current methods mostly ignore this. To do better, they create a new model that remembers recent changes using a special neural network called a k-memory Markov variational autoencoder. Their tests show this method reconstructs the signal information more accurately, especially when there is limited data capacity for feedback. This means explicitly considering time patterns helps compress CSI more effectively.
Channel State Information (CSI)Massive MIMOFDD SystemsNeural CompressionTemporal CorrelationMarkov ModelVariational AutoencoderLimited FeedbackLatent Dynamics
Authors
Efstathios Chatziloizos, Konstantinos Vandikas, Aneta Vulgarakis Feljan, Zheng Chen, Nikolaos Pappas
Abstract
This paper considers neural channel state information (CSI) compression for time-varying massive multiple-input multiple-output (MIMO) channels in frequency division duplex (FDD) systems with limited feedback resources. The main challenge lies in obtaining a compact and efficient representation of the CSI given that it exhibits strong temporal correlation across successive snapshots. Existing memoryless compression models do not exploit this property, while simple temporal extensions often incorporate multiple observations without explicitly modeling the latent dynamics. We propose a context-aware compression framework based on a k-memory Markov variational autoencoder (k-MMVAE), which uses a finite temporal window to capture the evolution of CSI in the latent space. The model introduces Markov-structured latent dynamics with finite memory, enabling efficient use of temporal dependencies for compression. Simulation results show that the proposed approach improves target CSI reconstruction performance compared to memoryless and weakly sequential baselines, particularly at low and moderate compression rates. These results suggest that explicit latent temporal modeling can provide an effective mechanism for CSI compression under limited feedback constraints.