Dataset

Data-centric AI; three criteria (quality, coverage, quantity); acquisition and annotation; data synthesis (rule-based, simulation, AI-powered, reverse instruction); model distillation; model collapse; data processing pipeline