Adversarial Learning on the Latent Space for Diverse Dialog Generation
Full Text: 2020.coling-main.441.pdfGenerating relevant responses in a dialog is challenging, and requires not only proper modeling of context in the conversation, but also being able to generate fluent sentences during inference. In this paper, we propose a two-step framework based on generative adversarial nets for generating conditioned responses. Our model first learns a meaningful representation of sentences by autoencoding, and then learns to map an input query to the response representation, which is in turn decoded as a response sentence. Both quantitative and qualitative evaluations show that our model generates more fluent, relevant, and diverse responses than existing state-of-the-art methods.
Citation
K. Khan, G. Sahu, V. Balasubramanian, L. Mou, O. Vechtomova. "Adversarial Learning on the Latent Space for Diverse Dialog Generation". Conference on Computational Linguistics (COLING), pp 5026-5034, December 2020.Keywords: | |
Category: | In Conference |
Web Links: | doi |
ACL |
BibTeX
@incollection{Khan+al:COLING20, author = {Kashif Khan and Gaurav Sahu and Vikash Balasubramanian and Lili Mou and Olga Vechtomova}, title = {Adversarial Learning on the Latent Space for Diverse Dialog Generation}, Pages = {5026-5034}, booktitle = {Conference on Computational Linguistics (COLING)}, year = 2020, }Last Updated: February 01, 2021
Submitted by Sabina P