Adapting Knowledge for Few-shot Table-to-Text Generation

arXiv:2302.12468v3 Announce Type: replace
Abstract: Pretrained language models (PLMs) have made remarkable progress in table-to-text generation tasks. However, the lack of domain-specific knowledge makes it challenging to bridge the topological gap between tabular data and text, especially in real-world applications with limited resources. To mitigate the limitation of insufficient labeled data, we propose a novel framework: Adapt-Knowledge-to-Generate (AKG). The core insight of AKG is to adapt unlabeled domain-specific knowledge into the model, which brings at least three benefits: (1) it …

abstract applications arxiv bridge cs.cl data domain few-shot gap however knowledge language language models progress resources table tabular tabular data tasks text text generation type world