AI RESEARCH

CHRONOS: Temporally-Aware Multi-Agent Coordination for Evolving Data Marketplaces

arXiv CS.AI

ArXi:2605.23887v1 Announce Type: cross Temporal knowledge-graph data marketplaces face three coupled failures in static designs: stale hybrid index shortcuts reduce recall as edges evolve, stationary Shapley pricing misattributes value after distribution shifts, and uncoordinated agents over-consume a shared differential-privacy budget. We present CHRONOS, a three-layer architecture providing a unified treatment of these challenges with explicit public and private separation.