UNO-DST: Leveraging Unlabelled Data in Zero-Shot Dialogue State Tracking

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Chuang, Zhang, Yan, Kan, Min-Yen, Li, Haizhou
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913296374824960
author Li, Chuang
Zhang, Yan
Kan, Min-Yen
Li, Haizhou
author_facet Li, Chuang
Zhang, Yan
Kan, Min-Yen
Li, Haizhou
contents Previous zero-shot dialogue state tracking (DST) methods only apply transfer learning, ignoring unlabelled data in the target domain. We transform zero-shot DST into few-shot DST by utilising such unlabelled data via joint and self-training methods. Our method incorporates auxiliary tasks that generate slot types as inverse prompts for main tasks, creating slot values during joint training. Cycle consistency between these two tasks enables the generation and selection of quality samples in unknown target domains for subsequent fine-tuning. This approach also facilitates automatic label creation, thereby optimizing the training and fine-tuning of DST models. We demonstrate this method's effectiveness on general language models in zero-shot scenarios, improving average joint goal accuracy by 8% across all domains in MultiWOZ.
format Preprint
id arxiv_https___arxiv_org_abs_2310_10492
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle UNO-DST: Leveraging Unlabelled Data in Zero-Shot Dialogue State Tracking
Li, Chuang
Zhang, Yan
Kan, Min-Yen
Li, Haizhou
Computation and Language
Previous zero-shot dialogue state tracking (DST) methods only apply transfer learning, ignoring unlabelled data in the target domain. We transform zero-shot DST into few-shot DST by utilising such unlabelled data via joint and self-training methods. Our method incorporates auxiliary tasks that generate slot types as inverse prompts for main tasks, creating slot values during joint training. Cycle consistency between these two tasks enables the generation and selection of quality samples in unknown target domains for subsequent fine-tuning. This approach also facilitates automatic label creation, thereby optimizing the training and fine-tuning of DST models. We demonstrate this method's effectiveness on general language models in zero-shot scenarios, improving average joint goal accuracy by 8% across all domains in MultiWOZ.
title UNO-DST: Leveraging Unlabelled Data in Zero-Shot Dialogue State Tracking
topic Computation and Language
url https://arxiv.org/abs/2310.10492