LG Uplus, Opt-AI Launch Token Optimization Research to Boost AI Operations Efficiency
TECHWORLD ·
✦ AI Summary
LG Uplus announced on the 2nd that it will push ahead with collaboration with Opt-AI.
The partnership aims to improve the operational efficiency of AI services and AI infrastructure by securing token optimization technology.
The two companies expanded their joint research scope to server GPU environments, and the amount of tokens that can be processed on the same GPU increased by up to 4 times compared with before.
LG Uplus announced on the 2nd that it will push ahead with collaboration with Opt-AI. The goal is to improve the operational efficiency of its AI services and AI infrastructure environment.
The partnership plan is focused on securing token optimization technology. The goal is to reduce GPU and power usage for operating AI services while improving response speed and service quality. Token optimization is a technology that lightens AI models and improves the efficiency of computational processes. Its effect lies in handling more requests with the same resources.
The collaboration structure combines the roles of LG Uplus and Opt-AI. LG Uplus will handle technology verification and application based on its experience operating actual AI services. Opt-AI will be responsible for research and development of lightweight, high-speed AI model execution technology.
The two companies have continued to cooperate in the on-device AI field. They also co-hosted the "Efficient AI Tech Seminar" on the 1st.
The seminar focused on sharing the latest research trends and industrial application cases in AI operations efficiency technologies. At the event, the companies discussed directions for technological advancement to improve AI operational efficiency. They also discussed future areas of cooperation.
Through this technology exchange, the two companies expanded the scope of their joint research to server GPU environments. Their key research plan is focused on making AI model computation processes more efficient so that, when operating large-scale AI services, more user requests can be processed with the same GPU resources.
Research on optimizing GPU-based AI models is currently under way, and initial results have also been produced. The two companies improved the computation process for generating AI model responses to better fit actual service environments.
As a result, the amount of tokens that can be processed on the same GPU has increased by up to 4 times compared with before. LG Uplus plans to apply the secured technology step by step to its AI services and large-scale AI infrastructure operating environment.
LG Uplus plans to continue improving AI operational efficiency. Lee Sang-yeop, CTO of LG Uplus, identified token processing efficiency per same resource as a key factor in AI competitiveness and said the company will expand research into token optimization technology in cooperation with various partners, including Opt-AI, and push ahead with efforts to create results from actual service applications.
Lee Jae-ho, CEO of Opt-AI, emphasized that in the generative AI era, securing operational efficiency along with high performance has become increasingly important. Based on that recognition, he said the company will work with Opt-AI and LG Uplus to advance AI optimization technology that can be verified in actual service environments, and through that contribute to improving the efficiency of the AI industry.
Source: TECHWORLD · Kim Hye-jin
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407717
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.