New system TYTAN uses LLMs to automate semantic schema construction from databases

Researchers have developed TYTAN, a system designed to automatically construct analytic semantic schemas from relational databases. This system combines symbolic database analysis with large language model (LLM) semantic inference to identify real-world entities, assign column roles, and generate names. When faced with ambiguity, TYTAN poses targeted natural-language questions to the user for clarification. Evaluations on eight databases demonstrated TYTAN's ability to achieve 100% coverage of entities and features, ensure 100% correctness in data retrieval instructions, and accurately characterize semantic roles with 92-100% agreement. AI

IMPACT Automates a critical knowledge-acquisition bottleneck in data analysis, potentially accelerating the scalability of analytic systems and empowering non-technical users.

RANK_REASON The cluster contains an academic paper detailing a new system and its evaluation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• Hugging Face
• LLM

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →