Self evolving ontology layer improves data agents working with diverse data
EvoOntology: A Self-Evolving Ontology Layer for Data Agents
Artificial IntelligenceComputation and LanguageDatabases
Summary
Data agents struggle to understand and use many different kinds of data because they only see limited information about it. The authors created EvoOntology, a special system that builds and updates a knowledge structure on its own to help agents find and use data better. This system keeps improving itself by checking and fixing its knowledge based on how well it helps the agents. Tests show that agents using EvoOntology do better at handling mixed data compared to older methods.
What this means in practice
- •For enterprise data teams: Integrate EvoOntology to improve automatic querying and interaction across various company data sources.
- •For business intelligence developers: Build smarter data tools that adapt their underlying knowledge structures dynamically for better handling of diverse data sets.
Authors
Meiduo Chong, Shaolei Zhang, Ju Fan, Xiaoyong Du
Abstract
Data agents aim to fulfill natural-language instructions over heterogeneous data, including tables, files, and databases. However, data agents face a challenging agent-data gap: heterogeneous data resides outside the agent, while the agent can access it (e.g., column names and file paths) only through generic tools. Existing approaches either let agents directly explore raw data sources or inject manually constructed semantic layers into prompts. However, neither scales well to large heterogeneous data sources nor adapts to different agent behaviors. In this paper, we introduce EvoOntology, a self-evolving ontology layer for data agents. EvoOntology encapsulates the ontology as an MCP server comprising a schema layer, a content layer, and a tool layer, enabling agents to actively query and interact with the ontology at runtime. To this end, we introduce a builder agent for autonomous ontology construction and a self-evolution loop that continuously refines the ontology through attribution-guided typed edits that are accepted only after a backbone-conditional paired evaluation. Experiments on three well-adopted data-agent benchmarks with four LLM backbones demonstrate that EvoOntology consistently outperforms strong baselines and existing semantic-layer approaches, effectively bridging the agent-data gap and enabling more effective interaction with heterogeneous data. Code: https://github.com/ruc-datalab/EvoOntology