Towards Faithfulness in Open Domain Table-to-text Generation from an Entity-centric View
This work addresses faithfulness issues in table-to-text generation for NLP applications, offering incremental improvements through entity-based approaches.
The paper tackled the problem of unfaithful text generation from tables by proposing entity-centric metrics to evaluate faithfulness and methods to improve it, resulting in improved generation fidelity in both full dataset and few-shot settings as shown by automatic and human evaluations.
In open domain table-to-text generation, we notice that the unfaithful generation usually contains hallucinated content which can not be aligned to any input table record. We thus try to evaluate the generation faithfulness with two entity-centric metrics: table record coverage and the ratio of hallucinated entities in text, both of which are shown to have strong agreement with human judgements. Then based on these metrics, we quantitatively analyze the correlation between training data quality and generation fidelity which indicates the potential usage of entity information in faithful generation. Motivated by these findings, we propose two methods for faithful generation: 1) augmented training by incorporating the auxiliary entity information, including both an augmented plan-based model and an unsupervised model and 2) training instance selection based on faithfulness ranking. We show these approaches improve generation fidelity in both full dataset setting and few shot learning settings by both automatic and human evaluations.