A new paper on arXiv introduces SCODEL, a shared canonical ontology designed to improve food entity linking in recipes. The authors note that identifying recipe entities and connecting them to ontology concepts is a fundamental step for semantic recipe understanding, but food mentions are often ambiguous, context-dependent, and lexically diverse, which complicates the task.
SCODEL aims to address these challenges by providing a unified set of ontology definitions that can be used to ground recipe entities. This allows for more consistent linking and supports the construction of ontology-grounded recipe knowledge graphs, which could enable richer semantic queries and reasoning over culinary data.
The paper is a single source, so there are no contrasting findings to compare. Its contribution appears to be the ontology itself and the framework for using it, though the abstract does not provide specific evaluation results or comparisons to existing methods.