TYTAN Automatically Constructs Semantic Schemas from Relational Data
August 7, 2026
TYTAN is a neurosymbolic system that automates the construction of analytic semantic schemas from relational databases. It combines symbolic database analysis with LLM-based inference to assign roles, propose entities, and resolve ambiguities through targeted user questioning.
HOW THIS AFFECTS YOU
●
builderYou can reduce the manual overhead of building semantic layers for data analysis tools.
●
designerThis enables more natural, conversational interfaces for complex database querying.