LG AI Research Institute is introducing 'Expert AI' tailored to solve problems in industrial settings. The initiative aims to create AI that works effectively in real-world environments by combining LG Group's industrial data and expertise.
At the 'LG AI Talk Concert 2026' held on September 14 at LG Science Park in Magok, Gangseo, Seoul, Im Woo-hyung, co-head of the LG AI Research Institute, stated, "The ultimate goal of Exaone is to develop Expert AI that addresses real-world issues." He added that the institute plans to continue improving the performance of its foundation models.
The LG AI Research Institute unveiled specialized Expert AI models for manufacturing, finance, and science. Im emphasized, "There are already numerous general-purpose AI models in the world, but AI must prove its effectiveness to be valued in industrial settings."
In the manufacturing sector, the institute showcased the achievements and future roadmap of 'Exaone Tabular,' which understands and predicts process conditions and quality information. LG Innotek applied Exaone Tabular to its manufacturing processes, reducing AI model response time from over 300 hours to just 50 hours, achieving an approximately 85% reduction in lead time.
In the science sector, the institute introduced 'Exaone Discovery,' which predicts the results of material design and synthesis. An 'AI Autonomous Laboratory' will begin operations by the end of this year, where AI will suggest experimental conditions, and robots will conduct experiments that AI will learn from.
In finance, the institute is expanding its market presence with 'Exaone Business Intelligence,' which performs data analysis, inference, prediction, and explanation. Lee Hwa-young, head of AI business development at LG AI Research Institute, noted that major domestic securities firms, large North American pension funds, top-tier financial institutions in the Middle East, and a hedge fund in the UK have already contracted with LG AI Research to develop financial AI products.
The LG AI Research Institute announced its commitment to maximizing the performance of the second model in the third phase evaluation of the 'Independent AI Foundation Model' project, promoted by the Ministry of Science and ICT.
Kim I-reun, head of the Exaone Lab at LG AI Research Institute, acknowledged that there are performance differences compared to global big tech models but emphasized that the focus of the third phase evaluation is to fully realize the potential of the second model.
He stated that the research is centered on enhancing data, pre-training, post-training, and reinforcement learning. The development is aimed at maximizing model performance as experienced by actual business users.
Kim added, "The second model is an intermediate step and a challenge toward becoming a frontier model, ultimately aiming to create a model that can be used in real industrial settings and achieve global frontier-level capabilities."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.