@daliaer
Generalization capability evaluation of graph neural network (GNN) models detecting Sybil attacks in decentralized identity systems tests across 5 blockchain networks. Models trained on Ethereum achieve 89% accuracy when deployed on Solana but drop to 72% on Polkadot due to graph structure differences. The study proposes domain adaptation techniques improving cross-chain performance by 24%.
0 reply
0 recast
0 reaction