OodGAN: Generative Adversarial Network for Out-of-Domain Data Generation
This addresses the expensive process of collecting annotated OOD data for dialog systems, offering a more efficient generation method, though it is incremental as it builds on existing GAN approaches.
The paper tackles the problem of generating out-of-domain (OOD) data for robust dialog systems by proposing OodGAN, a SeqGAN-based model that works directly on text, eliminating the need for auto-encoders. It results in a 67% relative improvement in FPR 0.95 on the ROSTD dataset and a 28% relative improvement on the OSQ dataset compared to state-of-the-art methods.
Detecting an Out-of-Domain (OOD) utterance is crucial for a robust dialog system. Most dialog systems are trained on a pool of annotated OOD data to achieve this goal. However, collecting the annotated OOD data for a given domain is an expensive process. To mitigate this issue, previous works have proposed generative adversarial networks (GAN) based models to generate OOD data for a given domain automatically. However, these proposed models do not work directly with the text. They work with the text's latent space instead, enforcing these models to include components responsible for encoding text into latent space and decoding it back, such as auto-encoder. These components increase the model complexity, making it difficult to train. We propose OodGAN, a sequential generative adversarial network (SeqGAN) based model for OOD data generation. Our proposed model works directly on the text and hence eliminates the need to include an auto-encoder. OOD data generated using OodGAN model outperforms state-of-the-art in OOD detection metrics for ROSTD (67% relative improvement in FPR 0.95) and OSQ datasets (28% relative improvement in FPR 0.95) (Zheng et al., 2020).