Collaborative Training of Gans in Continuous and Discrete Spaces for Text Generation
Applying generative adversarial networks (GANs) to text-related tasks is challenging due to the discrete nature of language. One line of research resolves this issue by employing reinforcement learning (RL) and optimizing the next-word sampling policy directly in a discrete action space. Such method...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9296209/ |
| _version_ | 1818917824679116800 |
|---|---|
| author | Yanghoon Kim Seungpil Won Seunghyun Yoon Kyomin Jung |
| author_facet | Yanghoon Kim Seungpil Won Seunghyun Yoon Kyomin Jung |
| author_sort | Yanghoon Kim |
| collection | DOAJ |
| description | Applying generative adversarial networks (GANs) to text-related tasks is challenging due to the discrete nature of language. One line of research resolves this issue by employing reinforcement learning (RL) and optimizing the next-word sampling policy directly in a discrete action space. Such methods compute the rewards from complete sentences and avoid error accumulation due to exposure bias. Other approaches employ approximation techniques that map the text to continuous representation in order to circumvent the non-differentiable discrete process. Particularly, autoencoder-based methods effectively produce robust representations that can model complex discrete structures. In this article, we propose a novel text GAN architecture that promotes the collaborative training of the continuous-space and discrete-space methods. Our method employs an autoencoder to learn an implicit data manifold, providing a learning objective for adversarial training in a continuous space. Furthermore, the complete textual output is directly evaluated and updated via RL in a discrete space. The collaborative interplay between the two adversarial trainings effectively regularize the text representations in different spaces. The experimental results on three standard benchmark datasets show that our model substantially outperforms state-of-the-art text GANs with respect to quality, diversity, and global consistency. |
| first_indexed | 2024-12-20T00:40:13Z |
| format | Article |
| id | doaj.art-d82f081423c34e1c918dfdddafac96dc |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-20T00:40:13Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-d82f081423c34e1c918dfdddafac96dc2022-12-21T19:59:38ZengIEEEIEEE Access2169-35362020-01-01822651522652310.1109/ACCESS.2020.30451669296209Collaborative Training of Gans in Continuous and Discrete Spaces for Text GenerationYanghoon Kim0https://orcid.org/0000-0002-9236-0702Seungpil Won1https://orcid.org/0000-0002-3557-4157Seunghyun Yoon2https://orcid.org/0000-0002-7262-3579Kyomin Jung3Department of Electrical and Computer Engineering, Seoul National University, Seoul, South KoreaDepartment of Electrical and Computer Engineering, Seoul National University, Seoul, South KoreaAdobe Research, San Jose, CA, USADepartment of Electrical and Computer Engineering, Seoul National University, Seoul, South KoreaApplying generative adversarial networks (GANs) to text-related tasks is challenging due to the discrete nature of language. One line of research resolves this issue by employing reinforcement learning (RL) and optimizing the next-word sampling policy directly in a discrete action space. Such methods compute the rewards from complete sentences and avoid error accumulation due to exposure bias. Other approaches employ approximation techniques that map the text to continuous representation in order to circumvent the non-differentiable discrete process. Particularly, autoencoder-based methods effectively produce robust representations that can model complex discrete structures. In this article, we propose a novel text GAN architecture that promotes the collaborative training of the continuous-space and discrete-space methods. Our method employs an autoencoder to learn an implicit data manifold, providing a learning objective for adversarial training in a continuous space. Furthermore, the complete textual output is directly evaluated and updated via RL in a discrete space. The collaborative interplay between the two adversarial trainings effectively regularize the text representations in different spaces. The experimental results on three standard benchmark datasets show that our model substantially outperforms state-of-the-art text GANs with respect to quality, diversity, and global consistency.https://ieeexplore.ieee.org/document/9296209/Adversarial trainingcollaborative trainingtext GAN |
| spellingShingle | Yanghoon Kim Seungpil Won Seunghyun Yoon Kyomin Jung Collaborative Training of Gans in Continuous and Discrete Spaces for Text Generation IEEE Access Adversarial training collaborative training text GAN |
| title | Collaborative Training of Gans in Continuous and Discrete Spaces for Text Generation |
| title_full | Collaborative Training of Gans in Continuous and Discrete Spaces for Text Generation |
| title_fullStr | Collaborative Training of Gans in Continuous and Discrete Spaces for Text Generation |
| title_full_unstemmed | Collaborative Training of Gans in Continuous and Discrete Spaces for Text Generation |
| title_short | Collaborative Training of Gans in Continuous and Discrete Spaces for Text Generation |
| title_sort | collaborative training of gans in continuous and discrete spaces for text generation |
| topic | Adversarial training collaborative training text GAN |
| url | https://ieeexplore.ieee.org/document/9296209/ |
| work_keys_str_mv | AT yanghoonkim collaborativetrainingofgansincontinuousanddiscretespacesfortextgeneration AT seungpilwon collaborativetrainingofgansincontinuousanddiscretespacesfortextgeneration AT seunghyunyoon collaborativetrainingofgansincontinuousanddiscretespacesfortextgeneration AT kyominjung collaborativetrainingofgansincontinuousanddiscretespacesfortextgeneration |