Metabuild develops 'Unfair Claim Detection AI Model' service

Metabuild, an artificial intelligence (AI) specialized company, announced on the 13th that it has recently developed an "AI model for detecting unfair claims." This development follows the development of an AI nursing insurance supply and demand forecasting model.

The "Unfair Claim Detection AI Model" was commercialized as an AI platform (MAI AUTO) based on new machine learning (MLOps) technology for the first time in Korea. "MAI AUTO" allows experts and field users to directly collect learning data and develop processing-learning-tuning-visualization-version management AI services of the entire AI learning process based on complete tools.

The 'Unfair Claims Detection AI Model' can increase accuracy and efficiency in responding to a new type of unfair detection, so overseas exports are also expected.

The system provides continuous collection, management, and monitoring of raw data and original data to detect unfair claims. Reliability and accuracy can be maintained with the new AI model distribution function retrained with new data.

Metabuild's "MAI AUTO" provides key technologies for fine tuning and natural language processing along with pre-learning language models that apply large language models (LLM) such as BERT, BART, and GPT. Based on the deep learning model desired by the user, it provides a domain-specific language model that has learned several knowledge in advance, and supports the creation of a model that the user wants by additionally learning a small amount of domain-specific learning data.

Metabuild is also providing AI counseling service, AI music composition service, and original text data learning, summary, and generation service based on MAI AUTO, in conjunction with digital human service (R2MIX). The company plans to provide agent-automated AI development service on its MESIM ESB and APIG product.​ 

Published: 13 Aug, 2023