- InnoCORE-LLM
- Platform for Real-world Innovation in Smart Manufacturing through AI (PRISM-AI)
- AI-CRED Institute(AI Co-Research & Education for Innovative Drug)
- AI-Transformed Aerospace Research Center
- Photonic artificial Intelligence COmputing REsearch Center
- Regenerative Medical Research Institute (reMRI) for Aging-related Diseases
- T-RAX: Transformative Railway with AX
- 5D AI-RI: 5D AI Robotics Initiative
- Artificial Cell Initiative
- POwer-autonomous Subsea Engineering for Intelligent Data-center Operating Networks
- Sustainable Materials Institute for Post-AI Era
- AI Meta-Scientist
InnoCORE-LLM
KEYWORDKeyword
OBJECTIVE Objectives
We aim to elevate the nation's artificial intelligence technology to a new level and dramatically enhance industrial applicability
and social acceptability by achieving the five-stage objectives of reasoning reinforcement, domain specialization,
multimodal expansion, reliability assurance, and integrated evaluation, centered on the four core competencies
of next-generation large language models: advanced reasoning, domain specialization,
multimodal expansion, and reliability assurance.
INTRODUCTION Director's Message
Hello. I am Jaegul Choo, Director of InnoCORE-LLM.
The research group has until recently focused its efforts primarily on developing science and technology-specialized language models.
Going forward, we plan to continuously strengthen multimodal functionality that can comprehend all non-textual elements
found in papers and reports, such as figures, diagrams, and mathematical expressions.
Ultimately, we plan to conduct research aimed at supporting the automation of the research process through what we call a Co-Scientist,
or researcher AI, an entity that aggregates and organizes multiple research findings at once and automatically formulates new hypotheses,
research directions, and experimental plans.
This means developing a researcher AI that aggregates, analyzes, and organizes multiple research results simultaneously,
identifies problems in existing studies, formulates new hypotheses, and automatically devises research and experimental plans
to support the research process.
The intention to integrate KONI into public-facing portals such as ScienceON and NTIS
in order to enhance AI-based analytical service functionality is rooted in the same vision.
Thank you.
IMPACT Impact
Technology Advancement
The transfer of core technologies and the demonstration of SaaS platform applications can accelerate the industrial adoption of AI technology
and contribute to productivity improvements among participating companies.
In addition, we seek to promote the diffusion of technology through open-source releases and patent acquisition,
while also realizing social value through public-sector utilization and improved access to information.
TALENT Ideal Talent
Talent who will build trustworthy next-generation AI with world-leading research capabilities and a spirit of convergent challenge.