Knowledge Graphs

How should enterprise facts be represented so that every answer can show where it came from?

Company data lives in systems, spreadsheets and email, with the same entity described in different ways. An answer built on it cannot be audited unless each fact keeps its source.

A graph represents entities and typed relations explicitly, and a provenance function attaches a source to each edge. Questions about impact and origin become path queries.

V are entities, R relation types, and π assigns each edge a source in S. The distance between two entities is the length of the shortest path that connects them.

1. Identify entities and relation types from the process.

2. Resolve duplicates into a single entity, keeping the evidence of the match.

3. Attach a source to every edge.

4. Answer impact questions as reachability queries.

1. Sources

2. Entity resolution

3. Graph with provenance

4. Path and impact queries

5. Answers with sources

No published experiments. The Enterprise Knowledge Graph lab runs shortest-path and downstream-impact queries on a synthetic graph.

Enterprise Knowledge Graph

No original research results are claimed. This area is presented as a technical perspective with a computable demonstration on synthetic data.

• Entity-resolution errors propagate into every query.
• A graph stores what was modeled; it does not discover missing relations.
• Maintaining provenance has a real engineering cost.

Background reading by other authors. These are not Menlo Tech publications, and no Menlo Tech paper is cited because none has been published.

• Hogan, A. et al. "Knowledge Graphs." ACM Computing Surveys 54(4), 2021.