English

LAD: Language Models as Data for Zero-Shot Dialog

Computation and Language 2022-08-01 v1 Artificial Intelligence

Abstract

To facilitate zero-shot generalization in taskoriented dialog, this paper proposes Language Models as Data (LAD). LAD is a paradigm for creating diverse and accurate synthetic data which conveys the necessary structural constraints and can be used to train a downstream neural dialog model. LAD leverages GPT-3 to induce linguistic diversity. LAD achieves significant performance gains in zero-shot settings on intent prediction (+15%), slot filling (+31.4 F-1) and next action prediction (+11 F1). Furthermore, an interactive human evaluation shows that training with LAD is competitive with training on human dialogs. LAD is open-sourced, with the code and data available at https://github.com/Shikib/lad.

Keywords

Cite

@article{arxiv.2207.14393,
  title  = {LAD: Language Models as Data for Zero-Shot Dialog},
  author = {Shikib Mehri and Yasemin Altun and Maxine Eskenazi},
  journal= {arXiv preprint arXiv:2207.14393},
  year   = {2022}
}

Comments

Accepted as a long paper to SIGDial 2022

R2 v1 2026-06-25T01:19:08.485Z