LAD: Language Models as Data for Zero-Shot Dialog
This addresses the challenge of data scarcity for zero-shot dialog systems, offering a competitive alternative to human-collected data, though it is incremental as it builds on existing language model capabilities.
The paper tackles the problem of zero-shot generalization in task-oriented dialog by proposing LAD, a paradigm for generating synthetic data using GPT-3, which results in performance gains of +15% in intent prediction, +31.4 F-1 in slot filling, and +11 F1 in next action prediction.
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.