Benchmarking Semantic Communications for Image Transmission Over MIMO Interference Channels

TL;DR

This work examines semantic communications over K-user MIMO interference channels and introduces an interference-robust semantic communication (IRSC) framework. By enabling CSI integration at the receiver (CSIR) or at both transmitter and receiver (CSITR) and training end-to-end neural transceivers with a dynamic, fairness-aware loss, IRSC learns to mitigate interference more effectively than CSI-free baselines, especially in low-SNR settings. The approach combines semantic encoding/decoding with joint source–channel coding and models the wireless channel within the network to enable end-to-end optimization. The results demonstrate substantial gains in image reconstruction quality (via SSIM) and robustness to interference, with CSIR offering practical advantages by avoiding CSI feedback.

Abstract

Semantic communications offer promising prospects for enhancing data transmission efficiency. However, existing schemes have predominantly concentrated on point-to-point transmissions. In this paper, we aim to investigate the validity of this claim in interference scenarios compared to baseline approaches. Specifically, our focus is on general multiple-input multiple-output (MIMO) interference channels, where we propose an interference-robust semantic communication (IRSC) scheme. This scheme involves the development of transceivers based on neural networks (NNs), which integrate channel state information (CSI) either solely at the receiver or at both transmitter and receiver ends. Moreover, we establish a composite loss function for training IRSC transceivers, along with a dynamic mechanism for updating the weights of various components in the loss function to enhance system fairness among users. Experimental results demonstrate that the proposed IRSC scheme effectively learns to mitigate interference and outperforms baseline approaches, particularly in low signal-to-noise (SNR) regimes.

Figures (5)

• Figure 1: Semantic communication system model over MIMO interference channels. The direct link is depicted by a solid line, and the interfering link is depicted by a dashed line.
• Figure 2: The proposed neural network structure for IRSC.
• Figure 3: Performance comparison of IRSC with other baseline schemes on the MNIST dataset and Fashion-MNIST dataset over MIMO interference channels.
• Figure 4: Comparison of reconstructed images produced by various schemes in the low SNR regime: An Example at SNR = $0$dB.
• Figure 5: SSIM comparisons versus the number of users over MIMO interference channels with test SNR of $0$dB and $15$dB.