How can dynamic reasoning over temporal knowledge graphs be optimized for low-latency applications?Answer not yet generated.