The Bangko Sentral ng Pilipinas (BSP) is exploring applications of machine learning (ML) techniques, particularly in the areas of natural language processing, nowcasting, and banking supervision.
“Central banks’ interest in ML has been increasing over the years, mainly due to its potential to enhance the existing tools used for regular monitoring as well as its ability to uncover underlying relationships between data to better understand the economy and the financial system,” BSP Governor Benjamin Diokno said.
Machine learning is a subfield of artificial intelligence (AI) involving algorithms that deliver output based on patterns learned from data.
Natural language processing at the BSP is used to convert text into data to produce a quantitative summary, such as the news sentiment index and economic policy uncertainty index that are currently being developed.
The BSP also employs ML approaches to generate nowcasts of regional inflation and domestic liquidity. These models supplement the BSP’s existing suite of models for macroeconomic forecasting.
In banking supervision, the BSP aims to utilize ML techniques to enhance its data validation processes and better identify atypical data.
Diokno said ML offers diverse opportunities in central banking, especially when combined with techniques from other disciplines, such as econometrics and network science.
Diokno also highlighted several challenges associated with ML processes. The most often-cited limitation is the black-box approach to ML, which could result in difficulties in interpreting the causal relationships in ML models.
Like traditional econometric techniques, ML algorithms may also encounter some challenges in accurately predicting tail risk or low likelihood events.
The adoption of ML models would also necessitate investments in information technology (IT) infrastructure and capacity building, as well as a change in the organizational mindset.
As the country builds a stronger and more technologically advanced Philippine economy, the BSP will continue to explore ML applications that can be useful in the conduct of its key functions, while carefully taking note of the associated challenges. (PR)
