A Data Enhancement Method for Non-Autoregressive Data Models Based on Joint Multi-Intent Detection and Slot Filling

The ability of Multi-Intent Spoken Language Understanding (SLU) to deal with multiple intentions has attracted increasing attention. This paper proposes a new approach to improve the accuracy of multi-intent detection and slot filling. In this research, we propose a new approach for data augmentation, which utilizes non-autoregressive models to enhance the performance of the task. Within this study, we demonstrate a substantial improvement over previous frameworks. Grounded on the results of our experiments with open data sets, we demonstrate that our approach is more effective than existing models.