GRMP crafts malicious updates via variational graph autoencoders on overheard benign feature graphs, degrading global LLM accuracy in federated IoA while evading statistical detection.
Zero-trust foundation models: A new paradigm for secure and collaborative artificial intelligence for internet of things,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NI 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Graph Representation-based Model Poisoning on the Heterogeneous Internet of Agents
GRMP crafts malicious updates via variational graph autoencoders on overheard benign feature graphs, degrading global LLM accuracy in federated IoA while evading statistical detection.