Verifiable Secure Aggregation via Dual Servers with Linear Tags in Federated Learning

This collection of recent academic papers covers security and privacy advances and threats across ML systems, edge/cloud cryptography, critical infrastructure, and agent/tool ecosystems. Key defensive contributions include a lightweight verifiable secure-aggregation scheme for federated learning using PRFs and a non-colluding dual-server design (low-cost tags and faster verification than OPSA), an attestation/allowlist extension to the Model Context Protocol (mcp-attested) for safer tool-server admission, and an RL-driven maintenance framework to adapt Android malware detectors under concept-d