Zero-Knowledge Proof of Training Secures Federated Learning Consensus and Privacy
The ZKPoT mechanism cryptographically validates model contributions using zk-SNARKs, resolving the critical trade-off between consensus efficiency and data privacy.
Epidemic Consensus Achieves Leaderless, Extreme-Scale, Collusion-Resistant Blockchain Finality
This new leaderless consensus protocol, Blockchain Epidemic Consensus Protocol (BECP), leverages viral communication to solve the decentralization-scalability trade-off.
Sparse Network Byzantine Agreement Achieves Near-Linear Fault Tolerance
A new fully-distributed protocol utilizes Byzantine Random Walks to achieve near-linear fault tolerance in sparse networks, fundamentally securing real-world peer-to-peer architectures.
Linear-Complexity Secret Sharing Unlocks Scalable Decentralized Randomness Beacons
A novel Publicly Verifiable Secret Sharing scheme reduces complexity to $O(n)$, enabling highly scalable, unbiasable randomness for large-scale consensus.
Angelfish Consensus Unifies Leader-DAG BFT for Optimal Throughput and Latency
The Angelfish hybrid consensus protocol dynamically integrates leader-based and DAG architectures, resolving the core BFT throughput-latency trade-off.
Set Byzantine Consensus Decouples Rollup Sequencing from Centralized Control
The research introduces Set Byzantine Consensus to construct a Decentralized Arranger, fundamentally solving rollup centralization by separating transaction content agreement from final ordering.
Set Byzantine Consensus Decentralizes Rollup Sequencing and Data Availability
Set Byzantine Consensus introduces a new primitive for L2s, enabling a decentralized 'arranger' service to eliminate sequencer centralization and censorship risk.
Graded Broadcast Protocol Decouples Agreement Achieving Optimal Asynchronous Latency
Falcon's Graded Broadcast bypasses mandatory agreement in ABFT, establishing a new primitive for lower latency and superior throughput in fully asynchronous systems.
Asymmetric DAG Consensus Unlocks Constant Finality with Local Trust Assumptions
By extending DAG-Rider to use asymmetric quorums, this work achieves constant-time BFT finality under realistic, locally defined trust models.
