Dunamu Machine Learning Team Achieves First Place in Financial Misinformation Detection Challenge (FMD)

Dunamu announced on the 23rd that its in-house Machine Learning (ML) team secured first place in the Financial Misinformation Detection (FMD) challenge at the International Conference on Computational Linguistics 'COLING 2025'.

'COLING' is one of the most prestigious academic conferences in the field of Natural Language Processing (NLP) and Computational Linguistics, where researchers and companies from around the world showcase cutting-edge technologies. The FMD challenge hosted at COLING 2025 focuses on tasks that evaluate the ability to detect misinformation in financial data and provide accurate analysis. It was established to address the severity of financial misinformation spreading through media.

The Dunamu Machine Learning team achieved first place in metrics such as F1, which measures the accuracy of misinformation detection, and ROUGE, which evaluates the rationale for misinformation discrimination, by utilizing Large Language Models (LLM). Notably, they achieved first place by directly modifying open-source large language models without using commercially available models like Chat GPT.

Hee-Soo Park, a researcher at Dunamu's Machine Learning team, stated, "We are honored that the passion and dedication of our team, which has worked towards investor protection and a transparent financial environment, have been recognized internationally." Dong-Jun Lee, the head of Dunamu's Machine Learning team, said, "This achievement proves our value of prioritizing customer trust, and we will continue to develop technologies and services that customers can rely on, contributing to market stability and transparency."