Committee-Based Byzantine Agreement Protocol Slashes Communication Complexity
A novel committee-based protocol achieves optimal asynchronous Byzantine agreement, drastically reducing cubic communication overhead.
Compositional Formal Verification Secures DAG Consensus Protocol Architectures
A new compositional framework using TLA+ achieves reusable formal verification for DAG consensus, halving proof effort and ensuring robust safety assurances for next-generation architectures.
Hybrid Ordering Multi-BFT Protocol Decouples Concurrency and Consistency
Orthrus introduces concurrent partial transaction ordering and a novel escrow mechanism to reduce consensus latency and maximize throughput in BFT systems.
Zero-Knowledge Proof of Training Secures Private Decentralized Federated Learning Consensus
ZKPoT introduces a zk-SNARK-based consensus mechanism that proves model accuracy without revealing private data, resolving the critical privacy-accuracy trade-off in decentralized AI.
Accountable Safety Decouples Liveness and Finality in Proof-of-Stake Consensus
This research introduces Accountable Safety, a new PoS property that guarantees finality or provides cryptographic proof of validator misbehavior under minimal synchrony.
Efficient Validated Agreement Bridges Complexity Gap for Secure State Replication
New signature-free validated Byzantine agreement protocols achieve optimal bit complexity, securing progress and external validity for high-performance state machine replication.
Asynchronous Verifiable Random Functions Achieve Optimal Leaderless BFT Consensus
AVRFs enable every node to verifiably compute the next proposer locally, eliminating leader election latency and achieving optimal asynchronous speed.
Constant-Cost Batch Verification for Private Computation over Secret-Shared Data
New silently verifiable proofs achieve constant-size verifier communication for batch ZKPs over secret shares, unlocking scalable private computation.
Zero-Knowledge Proof of Training Secures Private Federated Learning Consensus
ZKPoT consensus validates machine learning contributions privately using zk-SNARKs, balancing efficiency, security, and data privacy for decentralized AI.
