LG AI Research and Google Cloud have announced an expanded collaboration to develop LG’s latest EXAONE 3.0 generative AI models and the ChatEXAONE enterprise multimodal AI agent. This partnership builds on previous successful collaborations to advance AI capabilities.
- Previous Collaborations
- Past Work: EXAONE 1.0 (December 2021) and EXAONE 2.0 (July 2023).
- Achievements: Developed Korea’s first bilingual and multimodal generative AI model and subsequent advancements.
- EXAONE 3.0 and ChatEXAONE Developments
- Training Infrastructure: Utilized Google Cloud’s Cloud TPUs and GPUs via the AI Hypercomputer.
- Performance Gains: Reduced inference processing time by 56%, memory usage by 35%, and operating costs by 72% compared to the previous model.
- Capabilities: Bilingual model (Korean and English) with world-class performance in Korean.
- Technological and Operational Enhancements
- Data Utilized: 60 million specialized data sets, 350 million images, patents, software codes, and more.
- Efficiency: Achieved breakthroughs in scalability, performance, and cost efficiency due to Google Cloud’s infrastructure.
- Security and Data Governance
- Importance: Security is crucial for LG, dealing with major national technologies.
- Features: Google Cloud’s data governance and privacy controls, MLOps pipeline integration, and grounded data in LG’s corporate databases.
- Future Plans
- Google Cloud Marketplace: EXAONE 3.0 will be available on the marketplace.
- API Development: Exploring developer-friendly APIs to support enterprise AI agents.
LG AI Research’s expanded partnership with Google Cloud highlights the impact of collaborative innovation in AI. By leveraging Google Cloud’s advanced infrastructure, LG is poised to enhance its AI models’ performance and scalability, pushing the boundaries of AI technology.
