Variational Hierarchical Dialog Autoencoder for Dialog State Tracking Data Augmentation

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 3406–3425, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 3406 Variational Hierarchical Dialog Autoencoder for Dialog State Tracking Data Augmentation Kang Min Yoo1 Hanbit Lee1 Franck Dernoncourt2 Trung Bui2 Walter Chang2 Sang-goo Lee1 1Seoul National University, Seoul, Korea 2Adobe Research, San Jose, CA, USA {kangminyoo,skcheon,sglee}@europa.snu.ac.kr {dernonco,bui,wachang}@adobe.com Abstract Recent works have shown that generative data augmentation, where synthetic samples gen- erated from deep generative models comple- ment the training dataset, benefit NLP tasks. In this work, we extend this approach to the task of dialog state tracking for goal- oriented dialogs. Due to the inherent hierar- chical structure of goal-oriented dialogs over utterances and related annotations, the deep generative model must be capable of captur- ing the coherence among different hierarchies and types of dialog features. We propose the Variational Hierarchical Dialog Autoen- coder (VHDA) for modeling the complete as- pects of goal-oriented dialogs, including lin- guistic features and underlying structured an- notations, namely speaker information, dialog acts, and goals. The proposed architecture is designed to model each aspect of goal-oriented dialogs using inter-connected latent variables and learns to generate coherent goal-oriented dialogs from the latent spaces. To overcome training issues that arise from training com- plex variational models, we propose appropri- ate training strategies. Experiments on vari- ous dialog datasets show that our model im- proves the downstream dialog trackers’ robust- ness via generative data augmentation. We also discover additional benefits of our unified approach to modeling goal-oriented dialogs – dialog response generation and user simula- tion, where our model outperforms previous strong baselines. 1 Introduction Data augmentation, a technique that augments the training set with label-preserving synthetic sam- ples, is commonly employed in modern machine learning approaches. It has been used extensively in visual learning pipelines (Shorten and Khoshgof- taar, 2019) but less frequently for NLP tasks due to the lack of well-established techniques in the area. While some notable work exists in text classi- fication (Zhang et al., 2015), spoken language un- derstanding (Yoo et al., 2019), and machine trans- lation (Fadaee et al., 2017), we still lack the full understanding of utilizing generative models for text augmentation. Ideally, a data augmentation technique for super- vised tasks must synthesize distribution-preserving and sufficiently realistic samples. Current ap- proaches for data augmentation in NLP tasks mostly revolve around thesaurus data augmentation (Zhang et al., 2015), in which words that belong to the same semantic role are substituted with one another using a preconstructed lexicon, and noisy data augmentation (Wei and Zou, 2019) where ran- dom editing operations create perturbations in the language space. Thesaurus data augmentation re- quires a set of handcrafted semantic dictionaries, which are costly to build and maintain, whereas noisy data augmentation does not synthesize suf- ficiently realistic samples. The recent trend (Hu et al., 2017; Yoo et al., 2019; Shin et al., 2019) gravitates towards generative data augmentation (GDA), a class of techniques that leverage deep generative models such as VAEs to delegate …

Content truncated to 500 words