startthat.work home

PE-11 Knowledge graph construction from text (entity-relation graph; ontology population)

Convert resolved entities and extracted relations into a graph of typed nodes and edges, each linked to its source sentence, giving a graph with provenance as a draft result.

Prepare and extractEvidence: DocumentedPreparation technique, runs in step 5.2Strongest statement: describe

What it does

Converts extracted entities and relations into a graph of typed nodes and edges, each linked to its source sentence. Artefact: a graph with provenance. Read paths between entities and areas with few or contradictory edges.

When to use it

Use when you need to trace paths between people, organisations or concepts reported across many documents.

When it does not fit

Needs a fixed ontology and kept provenance. Extraction errors compound, and duplicates of one entity stay as separate nodes. No human-agreement figures for graph quality from text were found.

Result format

The result is a graph of typed nodes and edges, with each edge linked to the source sentence it came from. It names the paths between entities and the areas with few or contradictory edges.

Requirements on the material

PropertyRequirement typeRequirement
StructureApplies in some casesrequires=unstructured; requires=semi-structured
Content typeAlways appliesrequires=text
Unit of analysisAlways appliesrequires=document
Appraises source credibilityApplies in some casesno
Personal informationAlways appliesrequires=no personal information; requires=personal information

Not restricted on: Variable types, Time structure, How the sample was chosen, Level of detail available, Number of sources, Whether sources can be joined, Amount of material, Known data problems, Producer interest, Source dependence, Origin, Document genre, Language, Security classification, Licence.

Minimum materialA fixed ontology and a source sentence kept for every node and edge.
Additional sources neededreference structure
Not for (rules)none
Not for (text)No ontology is fixed and provenance is not kept, so extraction errors cannot be traced to the sentence.

Choosing and ordering

Kind of techniquePreparation technique, runs in step 5.2
Phase tagverify and prepare
Other phase tagsanalyse
Kind of findingpattern; gap
Question answeredwho; what
Strongest statement it supportsdescribe
Needs as input
Makes as output
Must exist firstNone
Loop targetreturn to verify and prepare

Library checks by stakes

Each stakes level includes the checks of the level below it. Checks that need a person apply at high stakes only, so the process leaves them out at lower stakes (decision 38). Repeatability: repeatable with fixed settings.

StakesCheckPerson involvedCounts as a different method
exploratoryspot checkoptionalno
decision-informingspot checkoptionalno
decision-informinghuman review sampleneeded; left out below high stakesno
decision-informingsource-span citationnoneno
high-stakesspot checkoptionalno
high-stakeshuman review sampleneededno
high-stakessource-span citationnoneno
high-stakesgold standard agreementconditionalno
high-stakesindependent reviewneededno
high-stakesprovenance lognoneno

Failure modes and limits

Model suitabilityhuman judgement required
Typical model failureKeeps duplicate nodes for one entity and drops the link from an edge to its source sentence.
Error sourcesinput data; data transformation; language model
Certainty limitsCertainty drops because extraction errors compound across steps (risk of bias), duplicate nodes remain (inconsistency), and no agreement figures exist for graph quality from text (indirectness).

Evidence and tag basis

Documented. The survey abstract covers creation, enrichment, quality assessment, refinement and publication, and does not mention extraction from text. (Hogan et al., Knowledge graphs, arXiv 2003.02320)

Tag basis: input profile inference; selection inference; sequencing inference; robustness inference. All tag values are inference unless a source is named.