Earnest AI Consortium Completes Phase 1 of National AGI Project, Launches Phase 2 with KAIST
Creatrip Team
2 months ago
Earnest AI-led consortium announced completion of Phase 1 of a government-funded next-generation AI (AGI) project, achieving key milestones toward causal-reasoning AGI. The team built a Dynamic Causal Knowledge Graph (DCKG) of about 2.83 million entries to capture cause-and-effect relationships rather than mere correlations, and published six papers at major AI conferences. They developed a neuro-symbolic pattern miner that extracts human-like causal inferences from text with over 86% accuracy (verified by experts), and established an automated AGI-Ops system to optimize AI learning and operation. Phase 2 will add Prof. Joo Jae-Geol’s lab from KAIST to advance an omni-modal (text, image, video) causal-knowledge reasoning engine. Earnest AI plans to integrate these core technologies into explainable AI (XAI) financial tools and other domain applications such as disaster response and healthcare.
