Samil PwC's Self-Developed AI Engine Ranks First on World's Largest AI Platform Evaluation

On August 22, Samil PwC announced that its self-developed artificial intelligence (AI) engine ranked first in the Korean language category of the "Search Model (MTEB, Massive Text Embedding Benchmark) Leaderboard" operated by Hugging Face, the world's largest machine learning platform.

The Korean embedding model developed by Samil PwC is considered an essential technology for generating accurate and meaningful answers by identifying the most contextually relevant information to a user's query. The model scored 77.01 points, surpassing models developed by global big tech companies such as Microsoft and Alibaba. Samil PwC stated, "This result signifies that Samil PwC has secured the most important core engine technology that determines the performance of AI services," adding, "It will serve as a key asset for differentiating Samil's AI services and creating revenue models in the future."

The Hugging Face MTEB Leaderboard is used as a key indicator for AI developers worldwide to objectively compare and evaluate the performance of text embedding models.

In addition, Lingo, a translation-specialized language model independently developed by Samil PwC, ranked first in the patent translation category at the Workshop on Machine Translation (WMT) in October last year.

Kim Sejun, Director of the AX Node GenAI Team at Samil PwC, said, "We will work to expand our Korean language embedding model to English and other language models, while also training it on internal data such as audit, tax, and deals, so that it can be widely applied to various AI services developed by Samil PwC."

This content was produced with the assistance of AI translation services.