[2024 Future Business Forum] Im Woohyeong, LG AI Research Institute Executive Director, "Using AI for Demand Forecasting and New Drug Development"

Major Affiliates Including Jeonja, Uplus, CNS, and Ensol
Pilot Deployment of Super Large Language Model 'ExaOne'
Utilization in Telecommunication Services, Product Development, and New Drug Development

LG Group is increasing productivity and reducing costs by applying AI to the production of TV products at major affiliates in electronics, chemicals, and telecommunication services, as well as to practical work in new drug development.

Im Woohyeong, Executive Director of LG AI Research, is participating in a panel discussion at the '2024 Asia Future Business Forum' hosted by Asia Economy on the 22nd at Lotte Hotel in Jung-gu, Seoul. Photo by Hyunmin Kim kimhyun81@

On the 22nd, Im Woo-hyung, Executive Director of LG AI Research Institute, stated at the 'Asia Future Enterprise Forum' held at Lotte Hotel in Sogong-dong, Jung-gu, Seoul, "We are utilizing LLMs (Large Language Models) such as ExaOne Universe and ExaOne Discovery in the manufacturing sites of our affiliates." The LG AI Research Institute is the organization responsible for overseeing AI research and education within LG Group.

LG Group is expanding AI utilization across all areas including research and development (R&D), product sales, production, and manufacturing. The three major platforms of LG AI Research Institute's ultra-large language model 'ExaOne'?'Universe,' 'Discovery,' and 'Atelier'?are being used in pilot forms at key affiliates such as LG Electronics, LG Uplus, LG CNS, and LG Energy Solution. This means AI is actively used in processes like enhancing TV picture quality and improving telecommunication service quality.

ExaOne Universe informs practitioners which internal company documents to refer to for making more accurate decisions. For example, when a customer center counselor uses ExaOne Universe to organize consultation content, it enables better customer service. ExaOne Discovery helps in effectively reading academic papers, patent documents, and so on. For instance, in chemical experiments, AI automatically extracts only the data usable for experiments from documents mixing chemical formulas, charts, and tables, and then conducts virtual experiments based on that data.

Executive Director Im explained, "We are introducing ExaOne in material development and vaccine new drug development," adding, "AI is also used to predict market demand during the R&D process." He continued, "AI handles both structured and unstructured data, and by utilizing this, market data can be understood more accurately," adding, "It helps to more precisely determine how much to adjust production volume and inventory."

Executive Director Im noted that no matter how good the technology is, it inevitably takes time for it to become widespread in the market. From the practitioner's perspective, there are parts where work methods need to change. Therefore, the process of understanding and persuading practitioners about how effective AI is is essential. He said, "While improving the completeness of AI technology is important, the most crucial part is persuading practitioners to recognize that using AI in the field helps their work," adding, "It is important to form a 'one team' between field practitioners and AI engineers to mutually understand AI knowledge, its strengths and weaknesses, and its necessity."