Global Model
A consolidated AI model trained on aggregated data from multiple sources, as opposed to localized or personalized models.
Definition
Centralized models that learn from all available data—often achieving high overall accuracy but potentially under-serving specific subpopulations. Global-model governance involves evaluating per-segment performance, assessing fairness across regions or demographics, and deciding when localized or personalized models (e.g., federated variants) are more appropriate for sensitive applications.
Real-World Example
A ride-hailing company deploys a global demand-forecasting model trained on city data worldwide. While it predicts London and New York trends well, it underperforms in emerging markets. The team then implements regional fine-tuning on local data to create hybrid models that blend global knowledge with local behavior patterns.
.svg)
