LG Uplus Wins Global Recognition for AI Optimization
As the axis of competition in artificial intelligence services shifts beyond model performance to data processing efficiency, LG Uplus has proven its AI technological prowess optimized for smartphone environments on the global stage.LG Uplus announced on Oct. 29 that two papers related to AI token o
LG Uplus has earned global recognition for its advancements in artificial intelligence (AI) technology, specifically its optimization for smartphone environments. This achievement comes after the company's research papers on AI token optimization were accepted for presentation at NeurIPS 2026, the world's leading AI conference. NeurIPS is a premier event attended by major tech companies, including Google, Meta, and OpenAI, and is set to take place in Sydney, Australia, beginning December 6.
At the heart of LG Uplus' research is the development of a lightweight, efficient language model designed for use in call services. By employing model lightweighting and inference acceleration technologies, the company has successfully optimized the model to operate within the constrained memory and power capacities of smartphones.
This optimization has resulted in a processing speed increase of approximately 4.5 times compared to the current standard, while preserving the model's performance equivalent to that of the most recent cloud AI systems.
A notable feature of this research is its emphasis on power efficiency. LG Uplus' approach has led to a reduction in battery consumption by around 77% compared to conventional methods. Additionally, the research introduces on-device learning technology, which enables AI training directly on the device, eliminating the need to transmit data to external servers.
This innovation not only enhances personal information protection but also allows for training at a level of accuracy comparable to server-based training, despite the limited computational resources of smartphones.
LG Uplus is currently evaluating the integration of these optimized capabilities into their service offerings. The company's ultimate goal is to leverage the verified smartphone environment technologies to enhance AI experiences in the broader server GPU landscape. Lee Sang-yeop, the Chief Technology Officer of LG Uplus, emphasized the significance of this recognition, stating that it validates their research in executing and training AI effectively even in constrained environments. He also hinted at future plans to expand the research scope to server environments.
Written by urgent.news from BusinessKorea's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
