Mind the Labels: Describing Relations in Knowledge Graphs With Pretrained Models
This addresses the issue of semantic inaccuracy in data-to-text generation for knowledge graphs, which is incremental as it builds on existing methods by emphasizing label diversity.
The paper tackles the problem of pretrained language models producing inaccurate outputs when data labels are ambiguous in data-to-text generation, specifically for describing relations between entities in knowledge graphs, and finds that models trained with diverse relation labels are robust in verbalizing 1,522 unique relations, including novel ones.
Pretrained language models (PLMs) for data-to-text (D2T) generation can use human-readable data labels such as column headings, keys, or relation names to generalize to out-of-domain examples. However, the models are well-known in producing semantically inaccurate outputs if these labels are ambiguous or incomplete, which is often the case in D2T datasets. In this paper, we expose this issue on the task of descibing a relation between two entities. For our experiments, we collect a novel dataset for verbalizing a diverse set of 1,522 unique relations from three large-scale knowledge graphs (Wikidata, DBPedia, YAGO). We find that although PLMs for D2T generation expectedly fail on unclear cases, models trained with a large variety of relation labels are surprisingly robust in verbalizing novel, unseen relations. We argue that using data with a diverse set of clear and meaningful labels is key to training D2T generation systems capable of generalizing to novel domains.