aiqbench analysis
FedLSG represents a significant advancement in securing Federated Graph Neural Networks by leveraging Large Language Models to interpret graph structures semantically rather than relying on brittle rule-based filters. This approach effectively neutralizes stealthy backdoor triggers while preserving the integrity of benign data during aggregation, addressing a critical vulnerability where traditional defenses often compromise model utility. The integration of LLMs into this specific security domain marks a clear evolution in how AI is applied to protect distributed machine learning systems from poisoning attacks.
FedLSG: LLM-Enhanced Semantic Calibration for Federated Graph Backdoor Defense
arXiv:2607.19674v1 Announce Type: cross Abstract: Federated Graph Neural Networks (FedGNNs) are highly vulnerable to backdoor poisoning, yet existing defenses typically rely on rule-based approaches that lack semantic understanding, making them vulnerable to stealthy triggers and harmful to benign structures. To solve this, we present FedLSG, the first framework that integrates large language models (LLMs) into federated graph backdoor defense. FedLSG introduces a graph and behavior to text grounding scheme that transforms local graph structures and client update behaviors into semantically rich natural language representations. The framework further adopts a lightweight student-teacher architecture. On the server side, a full scale LLM serves as a teacher, providing global contextual guidance and evaluating client updates during aggregation to identify potentially malicious participants. On the client side, a LoRA-based student is maintained to perform semantic reasoning, to suppress the influence of edges associated with backdoor triggers. By enabling semantic interpretation of both graph patterns and client behaviors, the framework adaptively incorporates rule-based signals into message passing and client aggregation for defense. Experiments demonstrate that FedLSG significantly improves resistance to backdoor attacks without compromising graph integrity.
Abstract:Federated Graph Neural Networks (FedGNNs) are highly vulnerable to backdoor poisoning, yet existing defenses typically rely on rule-based approaches that lack semantic understanding, making them vulnerable to stealthy triggers and harmful to benign structures. To solve this, we present FedLSG, the first framework that integrates large language models (LLMs) into federated graph backdoor defense. FedLSG introduces a graph and behavior to text grounding scheme that transforms local graph structures and client update behaviors into semantically rich natural language representations. The framework further adopts a lightweight student-teacher architecture. On the server side, a full scale LLM serves as a teacher, providing global contextual guidance and evaluating client updates during aggregation to identify potentially malicious participants. On the client side, a LoRA-based student is maintained to perform semantic reasoning, to suppress the influence of edges associated with backdoor triggers. By enabling semantic interpretation of both graph patterns and client behaviors, the framework adaptively incorporates rule-based signals into message passing and client aggregation for defense. Experiments demonstrate that FedLSG significantly improves resistance to backdoor attacks without compromising graph integrity.
Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2607.19674 [cs.CR]
(or arXiv:2607.19674v1 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2607.19674 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 277 words
Related coverage
News
Quantum Computing Report · Jul 25, 2026
Your Privacy My Cloak: Backdoor Attacks on Differentially Private Federated Learning
arXiv · Jun 16, 2026
Defending Against Backdoor Attacks via Alignment Checking in Model-Contrastive Federated Learning
arXiv · Jul 31, 2026
AI and Quantum Computing Briefing
livetradingnews.com · Aug 4, 2026