Crypto & Blockchain · Fri, Nov 13, 2026

Bitcoin, Ethereum, stablecoins, DeFi, smart contracts, scaling and consensus.

RSS feed
arXiv cs.AI · cs.MA

Cortex-FL: Agent-Mediated Collaboration for Federated Learning

In briefCortex-FL: An Agent-Based System for Coordinating Federated Learning Collaboration

SummaryThis paper presents Cortex-FL, a federated learning system in which agents handle collaboration. Users describe learning goals in natural language, and agents turn them into shared tasks, find collaborators, and negotiate participation agreements. Owners keep control of their local data, computation, and financial commitments. A four-node deployment demonstrates the full workflow, from collaboration formation through contract-governed training and model delivery to x402 settlement on a blockchain testnet.

Rui Sun, Jinfeng Peng, Zehua Cheng et al.

arXiv cs.DB · cs.DC

Omakase: Execution Congestion Control through Conflict-Graph Integration in Multi-Proposer Blockchain

In briefOmakase: Congestion control for blockchain transaction execution via conflict-graph integration

SummaryThis paper presents Omakase, a parallel execution framework for multi-proposer blockchains where transaction execution is the main bottleneck. Replicas pre-execute their own proposed blocks on idle cores, build per-block conflict graphs, and integrate them across a round with a greedy scheduler. A graph-aware optimistic concurrency control executor then runs the integrated graph in dependency order. On 15,000 Ethereum mainnet blocks and synthetic workloads, it achieves a 2.47x speedup over sequential execution and more stable throughput under high contention.

Dakai Kang, Ray Neiheiser, Alberto Sonnino et al.