A Scoping Review of Neurosymbolic Reasoning over Ontologies and Knowledge Graphs: System Categorization, Datasets, Evaluation Practices, and Open Challenges

Tracking #: 998-2034

Flag : New Paper Submitted

Authors: 

Julie Loesch
Michel Dumontier
Remzi Celebi

Submission Type: 

Survey

Full PDF Version: 

Supplementary Files: 

Cover Letter: 

RESUBMIT of 950-1975. Dear Editor, Please find enclosed the revised version of our manuscript, “A Scoping Review of Neurosymbolic Reasoning over Ontologies and Knowledge Graphs: System Categorization, Datasets, Evaluation Practices, and Open Challenges,” submitted for reconsideration to the Neurosymbolic Artificial Intelligence journal. We thank the Editor and the reviewers for their detailed and constructive feedback. We have carefully revised the manuscript in response to their comments and believe that the changes have substantially improved its clarity, organization, and contribution. The main revisions include an expanded introduction and background on knowledge graphs, ontologies, Description Logics, and reasoning tasks; a clearer positioning of the review relative to existing surveys; a substantially revised categorization of neurosymbolic systems, complemented by component-oriented architectural descriptions and diagrams; a reorganization of the Methods and Results sections; an expanded analysis of datasets, tasks, and evaluation metrics; and a revised evaluation framework highlighting key dimensions and limitations of current neurosymbolic evaluation practices. We also revised the dataset tables, including information on Description Logic expressivity where applicable, and addressed the reviewers’ detailed comments regarding terminology, system classification, and presentation. A detailed point-by-point response to all reviewer comments is provided in the supplementary files. For ease of review, we also provide a LaTeX-generated diff file highlighting the changes made between the original and revised manuscript. Kind regards, Julie Loesch on behalf of all authors

Previous Version: