References & Citations
Computer Science > Computation and Language
Title: LoFT: Enhancing Faithfulness and Diversity for Table-to-Text Generation via Logic Form Control
(Submitted on 6 Feb 2023)
Abstract: Logical Table-to-Text (LT2T) generation is tasked with generating logically faithful sentences from tables. There currently exists two challenges in the field: 1) Faithfulness: how to generate sentences that are factually correct given the table content; 2) Diversity: how to generate multiple sentences that offer different perspectives on the table. This work proposes LoFT, which utilizes logic forms as fact verifiers and content planners to control LT2T generation. Experimental results on the LogicNLG dataset demonstrate that LoFT is the first model that addresses unfaithfulness and lack of diversity issues simultaneously. Our code is publicly available at this https URL
Link back to: arXiv, form interface, contact.