LG AI Research Institute has introduced a specialized 'expert AI' tailored for the manufacturing, finance, and science sectors. This initiative aims to go beyond the competition of general artificial intelligence (AI) models and address complex challenges in industrial settings, thereby enhancing competitiveness.
On September 14, LG AI Research Institute held the 'LG AI Talk Concert 2026' at the LG Science Park in Magok, Gangseo-gu, Seoul, where it showcased the expert AI currently being developed and applied in manufacturing, finance, and science, along with its future business strategies.
Im Woo-hyung, co-head of LG AI Research Institute, stated in his keynote speech, "While creating good AI models is important, our mission is to solve difficult problems that have long plagued the industry." He added, "There are already numerous general AI models in the world that generate excellent responses, but to solve industrial problems, we need AI that can understand and make judgments based on countless variables and even 1% exceptions." He emphasized that AI must demonstrate tangible results to be recognized for its value in industrial settings.
During the event, LG AI Research Institute unveiled expert AI solutions developed over the past six years to address industrial challenges in manufacturing, science, and finance.
In the manufacturing sector, they presented 'EXAONE Tabular,' which understands and predicts process conditions and quality information, and 'EXAONE Omni-Inspect,' which can inspect products without needing retraining, even if the process or product appearance changes.
Lee Soon-young, head of the Data Intelligence Lab at LG AI Research Institute, and Yoo Jeong-sun from LG Innotek's AX division explained that utilizing industry-specific AI foundation models allows for rapid predictions of process outcomes with minimal on-site data and flexible responses to changing process conditions.
Looking ahead, LG AI Research Institute plans to enhance its technology with a 'vision inspection agent' that autonomously performs data sampling, labeling, and model training. Ultimately, they aim to implement an autonomous factory where the entire facility operates as a single intelligent entity, surpassing individual process automation.
In the science sector, they introduced 'EXAONE Discovery,' which is designed to predict the outcomes of material design and synthesis. The goal is to establish a system where AI conducts actual experiments, going beyond just design and prediction.
They also shared real-world applications, including a collaboration with LG Household & Health Care, where they reviewed 420,000 candidate substances and identified the hair loss treatment 'Rhamsydil' within a day. They are also developing next-generation oral peptide drugs with D&D Pharmatech and expanding the scope of new material discovery with GS Caltex, focusing on materials for next-generation AI data centers.
'EXAONE 4.5' has been integrated into the new drug review AI of the Ministry of Food and Drug Safety. LG AI Research Institute anticipates that AI will contribute to shortening the new drug approval process in the future.
In the finance sector, they are expanding their market presence with 'EXAONE Business Intelligence (BI),' which performs data analysis, inference, prediction, and explanation.
Lee Hwa-young, head of AI business development at LG AI Research Institute, noted, "We officially launched the finance-specific AI solution EXAONE BI earlier this year, which generates everything from data analysis to inference, prediction, and explanation through collaboration among multiple AI agents. We are expanding our business in the global market with the London Stock Exchange Group (LSEG) and in the domestic market with Koscom."
EXAONE BI incorporates the AI foundation model 'EXAONE Forecast,' which understands information that changes over time. LG AI Research Institute expects the application of explainable and reliable predictive AI in the financial market to expand across various asset classes, including bonds and commodities.
As the next step for expert AI, LG AI Research Institute is proposing AI that can make judgments and take actions in the real world. To achieve this, they are developing a 'robot foundation model' that serves as the brain for robots. They are also conducting research and development on proprietary algorithms focused on safety to prevent accidents that may occur during robot operations.
Im emphasized, "LG has been enabling AI to understand industrial problems for the past six years, and we are now preparing for AI that can make its own judgments and actions. We are moving toward a future of autonomous factories where the entire facility operates as a single intelligence, based on the robot foundation model."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.