The Financial Supervisory Service said on the 2nd that it is pushing to upgrade the trade analysis platform "VISTA" for unfair virtual asset trading investigations through artificial intelligence (AI). VISTA is a trade analysis system equipped with functions such as large-scale data analysis, automatic calculation of abnormal trading indicators, and visualization of trading patterns. It is currently being used as a key tool for analyzing and proving suspected transactions.
Financial Supervisory Service (FSS) in Dec. last year added two servers equipped with high-performance central processing units (CPUs) and graphics processing units (GPUs), laying the groundwork for applying AI algorithms based on shared computing resources and distributed data processing.
Financial Supervisory Service (FSS) used this to develop its own automatic suspected-interval extraction algorithm so staff can automatically detect suspected price-manipulation intervals that were identified manually through data analysis. The algorithm will be used to detect abnormal trading and crack down on suspected cases.
The algorithm partitions the suspect's transaction period into various detailed intervals ranging from a few seconds to several months and automatically calculates abnormal trading indicators for all possible intervals to perform an exhaustive search. This makes it possible to find every suspected interval in which price manipulation occurred, regardless of the suspect's number of manipulations or the period.
Financial Supervisory Service (FSS) plans to gradually expand AI analysis functions by the end of this year to detect unfair virtual asset transactions early and handle cases quickly. To that end, it has secured a server expansion budget of 170 million won this year and will use it to improve the performance of AI-related platforms.