Wednesday, Apr 23, 2025 | 03:00 PM - 03:45 PM
Location: Cedar Hall HQ1-1-660
.png)
OVERVIEW
Central bank communications is a vital policy tool to anchor expectations, build credibility, and enhance transparency. This presentation introduces a new AI tool by the IMF that classifies text to topic, forward-lookingness, sentiment, and audience. With these classifications, we build indicators that reveal monetary policy alignment, risk signaling, and consistency over time. Powered by a multilingual large language model, we apply our method to an unprecedented global dataset of over 70,000 documents and 20 million sentences from more than 160 countries spanning multiple languages. This AI tool enables large-scale benchmarking of communications practices across regions and levels of development, providing actionable insights for central banks. Already deployed in IMF technical assistance missions with Kazakhstan, Madagascar, Moldova, Sri Lanka, and Tonga, the AI tool is supporting central banks in refining monetary and financial stability messaging, identifying communications gaps, and strengthening public engagement. This work offers a compelling example of how AI can be meaningfully applied in central banking—translating advanced technology into tangible policy impact.
SPEAKERS
  |
||
Aubin Belalahy Second Vice Governor, Central Bank of Madagascar
|
||
Aubin Belalahy is the Second Deputy Governor of the Central Bank of Madagascar, a role he has held since May 2022, following more than two decades of service at the bank. He oversees monetary policy, monetary and foreign exchange operations, and research. An economist and engineer specialized in statistics, Aubin Belalahy holds several advanced degrees: he completed his higher education in diplomatic and strategic studies at the CEDS in Madagascar and obtained an ITB diploma from the Center for Banking Profession Training, France. Mr. Belalahy’s expertise extends to academia, with substantial experience teaching statistics and econometrics at the University of Antananarivo. He is also an active member of the academic community, contributing to various studies and macroeconomic modeling projects.
 |
||
Kei Moriya Senior Economic Systems Engineer, Information Technology Department, IMF
|
||
Kei Moriya is a Senior Economic Systems Engineer in the IT Department at the IMF. He works on data collection/management, automation, and applications of machine learning and economic modeling. He has worked on various projects at the IMF including forecasting using machine learning, nowcasting, and multiple projects related to central bank operations such as liquidity forecasting and central bank balance sheet modeling.
Prior to joining the IMF in 2016, Kei worked as a physicist in experimental nuclear and particle physics. He holds a Ph.D. in physics from Carnegie Mellon University and is originally from Japan.
|
||
Thiago Silva Senior Financial Sector Expert, Monetary & Capital Markets Department, IMF |
||
Thiago Christiano Silva is a Senior Financial Sector Expert in the Central Bank Operations Division of the Monetary and Capital Markets Department at the International Monetary Fund. His work spans central bank operations, governance, and financial sector analysis, with a focus on applying data science and machine learning techniques—including natural language processing and network analysis—to economic and financial data. He has been at the forefront of applying NLP to central bank communications, supporting technical assistance missions and contributing to the IMF’s Central Bank Transparency Code. Thiago has been using artificial intelligence to power much of this applied work, from processing large volumes of policy documents to generating insights that inform communication strategies and institutional assessments. Prior to joining the IMF in 2024, he served as a Senior Researcher and Head of the Financial Stability Division within the Research Department at the Central Bank of Brazil. He holds a Ph.D. in Computer Science (Artificial Intelligence) from the University of São Paulo and has published extensively in the intersection of economics, finance, and machine learning.
