Long Text Generation via Adversarial Training with Leaked Information

By Jiaxian Guo, Sidi Lu, Han Cai, Weinan Zhang, Yong Yu, Jun Wang. 2017

532 citations recorded by the source corpus.

Automatically generating coherent and semantically meaningful text has many applications in machine translation, dialogue systems, image captioning, etc. Recently, by combining with policy gradient, Generative Adversarial Nets(GAN) that use a discriminative model to guide the training of the generative model as a reinforcement learning policy has shown promising results in text generation. However, the scalar guiding signal is only available after the entire text has been generated and lacks intermediate information about text structure during the generative process. As such, it limits its success when the length of the generated text samples is long (more than 20 words). In this paper, we propose a new framework, called LeakGAN, to address the problem for long text generation. We allow the discriminative net to leak its own high-level extracted features to the generative net to further help the guidance. The generator incorporates such informative signals into all generation steps through an additional MANAGER module, which takes the extracted features of current generated words and outputs a latent vector to guide the WORKER module for next-word generation.Our extensive experiments on synthetic data and various real-world tasks with Turing test demonstrate that LeakGAN is highly effective in long text generation and also improves the performance in short text generation scenarios. More importantly, without any supervision, LeakGAN would be able to implicitly learn sentence structures only through the interaction between MANAGER and WORKER.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Generative Models

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Generative Models

Citations in the corpus

ColdGANs: Taming Language GANs with Cautious Sampling Strategies (2020, methodology): “Following [7, 21, 39, 15, 5, 23], we used the EMNLP2017 news dataset.”
Generative Adversarial Networks (2021, neutral): “By providing a thorough overview of GANs, we hope to make this complex and challenging topic more accessible to researchers and practitioners in the field, and inspire new innovations and applications.”
Towards Coherent and Cohesive Long-form Text Generation (2018, background): “, 2014) to incorporate feedback signals for text generation (Yu et al., 2017; Lin et al., 2017; Zhang et al., 2017; Guo et al., 2017; Fedus et al., 2018; Zhang et al., 2018).”
CausaLM: Causal Model Explanation Through Counterfactual Language Models (2020, background): “While in some cases, such as those presented in their paper, it might be plausible to generate counterfactual examples, in most cases in NLP it is still too hard to generate realistic texts with conditional generative models (Lin et al. 2017; Che et al. 2017; Rajeswar et al. 2017; Guo et al. 2018).”
Generating Informative and Diverse Conversational Responses via Adversarial Information Maximization (2018, background): “Adversarial learning [6, 26] has been shown to be successful in dialog generation, translation, image captioning and a series of natural language generation tasks [7, 12, 27, 28, 29, 30, 31, 32, 33, 34].”
Texygen: A Benchmarking Platform for Text Generation Models (2018, methodology): “LeakGAN [5] is a hierarchical reinforcement learning framework with two modules called Manager and Worker respectively.”
Language GANs Falling Short (2018, methodology): “, 2017), leaky discriminator (Guo et al., 2017), step-level loss instead of sequence level (Fedus et al.”
On Accurate Evaluation of GANs for Language Generation (2018, methodology): “…the LM; (v) We do not observe improvements of models with access to the discriminator’s state, suggesting that the previously reported good result [14] may be due to the RL setup; (vi) Supervised pretraining of the generator is extremely important, since training of every GAN model that achieves…”
Training language GANs from Scratch (2019, methodology): “We use EMNLP2017 News to compare with prior work[15, 20] but note that this dataset has limitations: a small vocabulary (5.”
Survey on reinforcement learning for language processing (2021, methodology): “2017) and LeakGAN (Guo et al.”
A Survey of the Usages of Deep Learning for Natural Language Processing (2020, background): “GANs have seen increasing use in text generation recently [219], [220].”
An Introduction to Image Synthesis with Generative Adversarial Nets (2018, unknown): “Research of applying GAN in natural language processing is also a growing trend, such as text modeling [13] [22] [23], dialogue generation [24], question answering [25] and neural machine translation [26].”
On Fenchel Mini-Max Learning (2019, background): “The following two benchmark datasets are considered: ( i ) EMNLP WMT news [26] and ( ii ) MS COCO [43].”
Integrating Physics-Based Modeling with Machine Learning: A Survey (2020, background): “Generative ML models have found tremendous success in areas such as speech recognition and generation [187], image generation [63], and natural language processing [100].”
Synthetic Data for Deep Learning (2019, methodology): “• training with the REINFORCE algorithm and other techniques from reinforcement learning that are able to handle discrete outputs; this path has been taken in the pioneering model named SeqGAN [678], LeakGAN for generating long text fragments [219], and MaskGAN that learns to fill in missing text wi”
Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Models (2023, background): “However, these methods, which create chains autoregressively in a single step, often suffer from error propagation as the number of steps increases (Guo et al., 2018; Chen et al., 2023b), due to compound errors.”
Feedback Adversarial Learning: Spatial Feedback for Improving Generative Adversarial Networks (2019, background): “GANs have continued to show impressive results in various applications such as image generation [46, 15, 63, 8], text to image synthesis [47, 26], future frame prediction [40, 55], image editing [67], novel view synthesis [52, 43], domain adaptation [7, 51], 3D modeling [59, 27], video generation [6”