arXiv:2611.12114v1 Announce Type: new
Abstract: Full-pipeline data analysis over real-world data lakes requires LLM agents to locate and integrate information across large collections of tables with different schemas. Existing data agents follow task-specific exploration, which exhausts the context window in massive lakes. As the context fills, the agent's reasoning degrades, leading it to locate and process incorrect data files. These errors propagate into the final answer. To overcome these limitations, we introduce Ontology-Grounded Data Agent (OGDA), which utilizes a compact ontology to reduce exploration overhead and constrain data processing. OGDA constructs the ontology through agentic hypothesis testing. At each step, a proposer agent induces schemas, concepts, and relations from raw sources, and a verifier checks each proposal against the underlying data. During task execution, the ontology serves as a representation of the data lake for the agent to explore. Every data-processing action is constrained by a task graph, which the agent can extend only after observing the semantic meaning of the data. Across KramaBench and DiscoveryBench with three LLM backbones, OGDA outperforms the strongest baselines by up to 10.82 points. These results suggest that explicitly modeling the data lake is a key step toward reliable end-to-end data agents.
arxiv
OGDA: Ontology-Grounded Data Agents for Reliable Analysis over Heterogeneous Data Lakes
Researchers introduce OGDA, a data agent that builds a compact ontology of a data lake through agentic hypothesis testing to reduce exploration overhead and constrain data processing. Proposer and verifier agents induce and check schemas, concepts, and relations, while a task graph limits data-processing actions until the semantic meaning of the data is observed. Across KramaBench and DiscoveryBench with three LLM backbones, OGDA outperforms the strongest baselines by up to 10.82 points.
It offers a concrete method for improving LLM data-agent reliability and context efficiency on large heterogeneous data lakes.
Comments & Questions (0)
No comments yet
Be the first to comment!