LG AI Research and Google Cloud have joined to develop LG's EXAONE 3.0 generative AI models and ChatEXAONE.
"Our collaboration with LG AI Research exemplifies the power of partnership in pushing the boundaries of AI innovation. By providing LG with our cutting-edge AI infrastructure, we're not only accelerating their groundbreaking work but also shaping the future of industries worldwide," said Ki-sung Chi, country director of Google Cloud Korea.
Heightened efficiency
EXAONE 3.0 models and a proprietary ChatEXAONE AI agent were trained using Google's Cloud TPUs and GPUs, which are available via Google Cloud's AI Hypercomputer. Google's AI accelerators claim to shorten the time for AI model training and heighten the process efficiency.
As a result, LG has significantly reduced inference processing time, memory usage, and operating costs, by 56%, 35%, and 72% respectively.
"EXAONE 3.0 has been trained on 60 million pieces of specialised data sets, 350 million images, patents, software codes, mathematics, and chemistry. Google Cloud's AI Hypercomputer has been instrumental in achieving LG AI Research's goals of AI model lightweight technology and cost efficiency," said Hwayoung (Edward) Lee, lead of AI Bis Dev and DX Group at the LG AI Research.
LG AI Research and Google Cloud previously worked closely to develop EXAONE 1.0 in 2021, Korea's first bilingual and multimodal generative AI model, and EXAONE 2.0 in 2023.
ChatEXAONE
LG also announced the opening of the beta service of ChatEXAONE, an enterprise AI agent based on EXAONE 3.0 for its affiliates' employees.
ChatEXAONE, LG's new enterprise AI agent, aim to revolutionise work productivity with its real-time web-based Q&A, document and image-based Q&A, coding, and database management features.
It has also adopted 'Retrieval-Augmented Generation' (RAG) technology, which can understand prompt contexts and respond with updated information.
