@inproceedings{cai-etal-2022-advancing,
    title = "Advancing Semi-Supervised Task Oriented Dialog Systems by {JSA} Learning of Discrete Latent Variable Models",
    author = "Cai, Yucheng  and
      Liu, Hong  and
      Ou, Zhijian  and
      Huang, Yi  and
      Feng, Junlan",
    editor = "Lemon, Oliver  and
      Hakkani-Tur, Dilek  and
      Li, Junyi Jessy  and
      Ashrafzadeh, Arash  and
      Garcia, Daniel Hern{\'a}ndez  and
      Alikhani, Malihe  and
      Vandyke, David  and
      Du{\v{s}}ek, Ond{\v{r}}ej",
    booktitle = "Proceedings of the 23rd Annual Meeting of the Special Interest Group on Discourse and Dialogue",
    month = sep,
    year = "2022",
    address = "Edinburgh, UK",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/add-emnlp-2024-awards/2022.sigdial-1.44/",
    doi = "10.18653/v1/2022.sigdial-1.44",
    pages = "456--467",
    abstract = "Developing semi-supervised task-oriented dialog (TOD) systems by leveraging unlabeled dialog data has attracted increasing interests. For semi-supervised learning of latent state TOD models, variational learning is often used, but suffers from the annoying high-variance of the gradients propagated through discrete latent variables and the drawback of indirectly optimizing the target log-likelihood. Recently, an alternative algorithm, called joint stochastic approximation (JSA), has emerged for learning discrete latent variable models with impressive performances. In this paper, we propose to apply JSA to semi-supervised learning of the latent state TOD models, which is referred to as JSA-TOD. To our knowledge, JSA-TOD represents the first work in developing JSA based semi-supervised learning of discrete latent variable conditional models for such long sequential generation problems like in TOD systems. Extensive experiments show that JSA-TOD significantly outperforms its variational learning counterpart. Remarkably, semi-supervised JSA-TOD using 20{\%} labels performs close to the full-supervised baseline on MultiWOZ2.1."
}