[MA] Evaluating the Impact of Comments in Knowledge Graphs on Question Answering Augmented with Language Models
Description:
Previous work has shown that changing the representation of knowledge graphs impacts the F1 score of question answering [1].
However, adding comments to terms of an ontology did not improve results in baseline experiments.
Your task is
(i) to distinguish variants of comments in a knowledge graph, considering different RDF terms and learnings from research fields like linguistics,
(ii) to draft evaluation data for the variants, building on an existing knowledge graph [2], and
(iii) to conduct experiments with a given language model-augmented question answering agent to compare F1 scores between variants.
Please note that providing coherent reasoning for your choices in solving the given task is required.
We prefer students with a good grade in Foundations of Linked Data.
We expect an initial exposé that outlines the planned approach and some related work and participation in monthly online update meetings.
If you are interested contact Daniel Henselmann.
[1] Henselmann, D., Dorsch, R., Harth, A., 2025. Impact of Knowledge Graph Representations on Question Answering with Language Models, in: Advanced Information Systems Engineering Workshops, Lecture Notes in Business Information Processing. Presented at the CAiSE 2025, Springer Cham, Vienna, Austria, pp. 81–92. https://doi.org/10.1007/978-3-031-94931-9_7