Wireless Backdoor Attack and Defense for Semantic Communications over Multiple Access Channel

2026-06-29

AI Summary

This work addresses the threat of selective wireless backdoor attacks in shared-access semantic communication systems operating over multi-access channels. To this end, it proposes a low-power over-the-air trigger waveform injection method capable of covertly manipulating the semantic inference outcomes of specific users during the training phase. The study introduces, for the first time, a selective over-the-air backdoor attack model tailored to multi-user semantic communications and develops a trigger-aware robust training framework. This defense mechanism enables accurate semantic recovery under poisoned observations by jointly optimizing semantic encoding, reconstruction, and classification. Experimental results demonstrate that the proposed approach effectively mitigates targeted backdoor attacks while preserving high semantic accuracy and significantly enhancing system security.

Abstract

Semantic communication (SemCom) aims to preserve semantic meaning and task-oriented information beyond conventional message recovery over wireless channels. The adoption of SemCom in shared-access wireless networks introduces new vulnerabilities for multi-user semantic inference. This paper considers a SemCom system for two transmitters communicating with a common receiver over a multiple access channel. Each transmitter maps source information into latent semantic representations, while the receiver jointly reconstructs and classifies the semantic information for both transmitters. A selective over-the-air backdoor (Trojan) attack is presented in which an adversary transmits a low-power trigger waveform over the air and injects it into the shared received signal during training. By transmitting the trigger again during testing, this stealthy, low-power attack selectively manipulates the semantic inference for one transmitter while minimally affecting the inference of the other transmitter. To mitigate this vulnerability, a trigger-aware defense mechanism is developed to preserve correct semantic labels under trigger-contaminated wireless observations. The results demonstrate both the vulnerability of shared-access SemCom systems to selective over-the-air backdoor attacks and the effectiveness of trigger-aware robust training for semantic protection.

Problem

• Semantic Communication
• Backdoor Attack
• Multiple Access Channel
• Wireless Security
• Adversarial Trigger

Innovation

• semantic communication
• backdoor attack
• multiple access channel
• trigger-aware defense
• over-the-air injection