TempLM: Distilling Language Models into Template-Based Generators
Tianyi Zhang, Mina Lee, Xiang Lisa Li, Ende Shen, Tatsunori Hashimoto
Abstract
While pretrained language models (PLMs) have greatly improved text generation, they have also been known to produce unfaithful or inappropriate content. In contrast, classic template-based systems provide strong guarantees of faithfulness at the cost of fluency. We propose TempLM, which achieves the best of both worlds by distilling a PLM into a template-based generator. On the E2E and SynthBio data-to-text datasets, we show that TempLM is more faithful than the original PLM and is more fluent than prior template systems. Notably, on an out-of-domain evaluation, TempLM reduces a finetuned BART model’s unfaithfulness rate from 83% to 0%. In a human study, we find that TempLM’s templates substantially improve upon human-written ones in BERTScore.
BibTeX
@inproceedings{zhang-etal-2023-templm,
title = "{T}emp{LM}: Distilling Language Models into Template-Based Generators",
author = "Zhang, Tianyi and
Lee, Mina and
Li, Xiang Lisa and
Shen, Ende and
Hashimoto, Tatsunori",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2023",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-acl.124/",
doi = "10.18653/v1/2023.findings-acl.124",
pages = "1970--1994"
}