Working Papers

This paper proposes the use of synthetic training data generated by large language models
to improve machine learning SDG classifiers. It shows that supplementing existing training data with
synthetic data produced by the ChatGPT tool improves the performance of the SDGClassy classifier.
This addition of synthetic data is especially useful in building SDG classifiers given the limited availability
of properly labeled data and the complex, interconnected nature of the SDGs. Synthetic data thus enables
more effective machine-learning applications in this context.

Recent calls for more intense debate on and reforms to the international monetary system imply that the current system is unable to respond appropriately and adequately to challenges that have appeared, or become more acute, in recent years. This paper focuses on four such challenges: ensuring an orderly exit from global imbalances, facilitating more complementary adjustments between surplus and deficit countries without recessionary impacts, better supporting international trade by reducing currency volatility and better providing development and climate finance. After describing them, it proposes reforms to enable the international monetary system to better respond to these challenges…