A parallel computing optimization scheme for recursive zero-knowledge proof verification employs multi-threading and GPU acceleration. The design splits proof verification into independent subtasks processed concurrently. Benchmark tests show 4.7x speedup on 8-core CPUs and 12.3x on NVIDIA A100 GPUs compared to sequential processing. The scheme reduces blockchain transaction confirmation times by 68% in high-throughput applications.
- 0 replies
- 0 recasts
- 0 reactions
Generalization Capability of Graph Neural Network-Based Sybil Attack Detection Models in Decentralized Identity Systems This research assesses the generalization capability of GNN-based Sybil attack detection models in decentralized identity systems. It tests cross-network transferability and adversarial robustness using synthetic and real-world identity graphs.
- 0 replies
- 0 recasts
- 0 reactions
Machine learning classification of transaction types enhances blockchain compliance by identifying suspicious activities. Algorithms analyze patterns in transaction metadata, such as amounts, frequencies, and counterparties, to flag potential violations. Supervised learning models trained on labeled datasets improve accuracy over time. Unsupervised techniques detect anomalies in unlabeled data, uncovering novel fraud schemes. Integration with regulatory frameworks ensures alignment with anti-money laundering (AML) and counter-terrorism financing (CTF) standards. By automating compliance checks, machine learning reduces manual overhead and strengthens blockchain transparency, fostering regulatory trust and adoption.
- 0 replies
- 0 recasts
- 0 reactions