Zero-Knowledge Proof of Training Secures Decentralized Federated AI
A new Zero-Knowledge Proof of Training consensus leverages zk-SNARKs to cryptographically verify model accuracy without exposing private data, solving the fundamental privacy-accuracy trade-off in decentralized AI.
Eliminating Threshold Cryptography Latency in Byzantine Fault Tolerant Consensus
Foundational research eliminates the inherent one-message latency price of threshold cryptography in BFT systems, enabling faster, provably secure on-chain randomness.
Optimal DAG BFT Achieves Theoretical Minimum Latency
Mysticeti-C, a new DAG-BFT, achieves the theoretical three-round latency lower bound by eliminating block certification, unifying high throughput with instant finality.
Practical Asynchronous BFT Achieves Superior Performance through Designated Leader Pipelining
This protocol merges leader-driven efficiency with asynchronous resilience, creating a simple, high-performance BFT mechanism for real-world decentralized systems.
Empirical Analysis Exposes Secret Leader Election Vulnerability to Coordinated Attacks
New simulations reveal current Secret Leader Election protocols fail against coordinated validator group attacks, demanding novel cryptographic resilience.
Express Consensus Algorithm Accelerates Proof-of-Authority Blockchains via Compact Block Protocol
ExClique fundamentally enhances Proof-of-Authority throughput by compacting block data and eliminating consensus-disrupting no-turn blocks.
Proof of Useful Work Unifies Consensus Security and Verifiable Computation Marketplace
A novel Proof of Useful Work protocol embeds SNARK generation into consensus, solving energy waste and creating a decentralized verifiable computation market.
Cardano Network Partitioned by Legacy Delegation Transaction Flaw
A legacy software vulnerability allowed a malformed delegation transaction to partition the network, compromising chain integrity.
Homomorphic Encryption and VRF Achieve Scalable Unpredictable On-Chain Randomness
Homomorphic encryption combined with VRFs constructs a linear-scaling distributed randomness beacon, eliminating pre-computation bias in consensus leader selection.
