Denoising Pre-Training and Data Augmentation Strategies for Enhanced RDF Verbalization with Transformers
This work is significant for researchers and developers working with Knowledge Bases, as it improves the interpretability of abstract RDF data for humans by enhancing automated text generation.
This paper addresses the challenge of converting RDF triples into human-readable text, a process known as RDF verbalization. By employing denoising pre-training and data augmentation strategies with a Transformer model, the authors achieved significant improvements in BLEU scores: 3.73% for seen categories, 126.05% for unseen entities, and 88.16% for unseen categories.
The task of verbalization of RDF triples has known a growth in popularity due to the rising ubiquity of Knowledge Bases (KBs). The formalism of RDF triples is a simple and efficient way to store facts at a large scale. However, its abstract representation makes it difficult for humans to interpret. For this purpose, the WebNLG challenge aims at promoting automated RDF-to-text generation. We propose to leverage pre-trainings from augmented data with the Transformer model using a data augmentation strategy. Our experiment results show a minimum relative increases of 3.73%, 126.05% and 88.16% in BLEU score for seen categories, unseen entities and unseen categories respectively over the standard training.