State Value Generation with Prompt Learning and Self-Training for Low-Resource Dialogue State Tracking

TL;DR

This paper addresses low-resource dialogue state tracking by proposing SVAG, a State VAlue Generation framework that splits DST into state value generation and domain slot generation and leverages prompt learning. It introduces a self-training loop guided by a novel state value estimator that discriminates between correct, incomplete, and incorrect state-value generation, along with synthetic negative samples to mitigate error reinforcement. A prompt-based domain slot generator (with inverse prompts) further maps generated values to their corresponding slots, while belief state updating ensures coherence across turns. Experiments on MultiWOZ 2.1 show SVAG achieving state-of-the-art joint-goal accuracy under data ratios of 5%, 10%, and 25% with models under 100B parameters, and competitive results versus substantially larger models. Overall, SVAG demonstrates how targeted value-generation, estimator-based self-training, and prompt-based slot mapping can substantially improve performance in data-scarce, task-oriented dialogue settings.

Abstract

Recently, low-resource dialogue state tracking (DST) has received increasing attention. First obtaining state values then based on values to generate slot types has made great progress in this task. However, obtaining state values is still an under-studied problem. Existing extraction-based approaches cannot capture values that require the understanding of context and are not generalizable either. To address these issues, we propose a novel State VAlue Generation based framework (SVAG), decomposing DST into state value generation and domain slot generation. Specifically, we propose to generate state values and use self-training to further improve state value generation. Moreover, we design an estimator aiming at detecting incomplete generation and incorrect generation for pseudo-labeled data selection during self-training. Experimental results on the MultiWOZ 2.1 dataset show that our method which has only less than 1 billion parameters achieves state-of-the-art performance under the data ratio settings of 5%, 10%, and 25% when limited to models under 100 billion parameters. Compared to models with more than 100 billion parameters, SVAG still reaches competitive results.

Figures (4)

• Figure 1: Three main issues of state value generation in DST: (a) "don't care" should be generated, (b) "hotel" should be distinguished from "guesthouse", and (c) "centre" should be inferred from the first turn. Words in blue are state values that can be extracted directly.
• Figure 2: The overview of our proposed framework. There are three main components of our framework: a state value generator, a self-training strategy, and a domain slot generator. Given the dialogue history, the state value generator first generates the state values in the current turn, then the domain slot generator generates the slot type for each generated state value. Finally, we use the turn labels to update the belief state.
• Figure 3: The model architecture of our proposed state value estimator. Given the dialogue history and the generated state values, the model predicts whether all the state values are correctly generated. This figure shows an example that the estimator detects incomplete generation.
• Figure 4: An example of our proposed negative sampling.