Added Objective Guided Optimization of Adversarial Text Generation Method
Abstract
The problem of error accumulation is caused by the supervision of deep neural network text generation model. In order to solve this problem, a text generation model based on the reinforcement of antagonistic thought training is proposed.The adversarial network can be generated by the proposed model, and then the adversarial network can be used for identification, the learning reward function can be optimized, and the generated model can be optimized to reduce the probability of error accumulation.More text structure knowledge can be added into the generated text model by integrating the target guidance feature into the actual generation process to make the generated text model have higher authenticity. In this paper, the author optimizes the adversarial text generation method on the basis of target-guided optimization, which can be used for reference by practitioners.References
[1] ZHANG Y, ZHANG Y, ZHANG Y, et al. A novel image generation algorithm based on generative adversarial networks [J]. Journal of Computer Research and Development, 2016, 23 (1) : 1-8.
[2] HU MAo-han. Research on text generation based on generative adversity-network [D]. University of Electronic Science and Technology of China,2020.
[3] Wang Y, Wang Y, Wang Y, et al. A new approach to conditional text generation based on deep learning [J]. Huaqiao University,2020.
[4] Xu Cong, Li Qing, Zhang Dezheng, Chen Peng, Cui Jiarui. Journal of engineering
[5] Pang suppository. Research and Implementation of Decoy Document Generation Based on LeakGAN [D]. Beijing Jiaotong University,2019.
Authors contributing to this journal agree to publish their articles under the Creative Commons Attribution-Noncommercial 4.0 International License, allowing third parties to share their work (copy, distribute, transmit) and to adapt it, under the condition that the authors are given credit, that the work is not used for commercial purposes, and that in the event of reuse or distribution, the terms of this license are made clear. With this license, the authors hold the copyright without restrictions and are allowed to retain publishing rights without restrictions as long as this journal is the original publisher of the articles.
