Long Text Generation via Adversarial Training with Leaked Information

Abstract

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.

14 pages, AAAI 2018

References in corpus (7)

• Microsoft COCO Captions: Data Collection and Evaluation Server
• Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks
• Text Understanding from Scratch
• Maximum-Likelihood Augmented Discrete Generative Adversarial Networks
• Adversarial Ranking for Language Generation
• Adversarial Generation of Natural Language
• Data Generation as Sequential Decision Making

Cited by in corpus (47)

• A Note on the Inception Score
• An Introduction to Image Synthesis with Generative Adversarial Nets
• Texygen: A Benchmarking Platform for Text Generation Models
• Survey on reinforcement learning for language processing
• A survey on text generation using generative adversarial networks
• Language GANs Falling Short
• On Accurate Evaluation of GANs for Language Generation
• Topic-Guided Variational Autoencoders for Text Generation
• Neural Text Generation: Past, Present and Beyond
• Reinforcement Learning for Generative AI: State of the Art, Opportunities and Open Research Challenges
• Creativity and Machine Learning: A Survey
• Generative Adversarial Network Training is a Continual Learning Problem
• A Tutorial on Deep Latent Variable Models of Natural Language
• Knowledge-Enhanced Personalized Review Generation with Capsule Graph Neural Network
• ValueNet: A New Dataset for Human Value Driven Dialogue System
• Toward Diverse Text Generation with Inverse Reinforcement Learning
• Evaluating Text GANs as Language Models
• Learning Semantic Sentence Embeddings using Sequential Pair-wise Discriminator
• CoT: Cooperative Training for Generative Modeling of Discrete Data
• Paraphrase Generation with Deep Reinforcement Learning
• New Era of Deeplearning-Based Malware Intrusion Detection: The Malware Detection and Prediction Based On Deep Learning
• ColdGANs: Taming Language GANs with Cautious Sampling Strategies
• Summarize, Outline, and Elaborate: Long-Text Generation via Hierarchical Supervision from Extractive Summaries
• Rethinking Exposure Bias In Language Modeling
• CatGAN: Category-aware Generative Adversarial Networks with Hierarchical Evolutionary Learning for Category Text Generation
• Reward Constrained Interactive Recommendation with Natural Language Feedback
• Sequence Generation with Guider Network
• Extractive Summary as Discrete Latent Variables
• Revision in Continuous Space: Unsupervised Text Style Transfer without Adversarial Learning
• Conditional Hybrid GAN for Sequence Generation
• Abstractive Text Summarization by Incorporating Reader Comments
• Meta-CoTGAN: A Meta Cooperative Training Paradigm for Improving Adversarial Text Generation
• Controllable and Diverse Text Generation in E-commerce
• Adversarial Sub-sequence for Text Generation
• Towards Coherent and Cohesive Long-form Text Generation
• The Image Local Autoregressive Transformer
• Towards Controllable and Personalized Review Generation
• Protecting Anonymous Speech: A Generative Adversarial Network Methodology for Removing Stylistic Indicators in Text
• Multimodal Story Generation on Plural Images
• Claim Verification using a Multi-GAN based Model
• Generating Symbolic Reasoning Problems with Transformer GANs
• Improvement in Machine Translation with Generative Adversarial Networks
• Improving Adversarial Text Generation by Modeling the Distant Future
• Generation of Synthetic Electronic Medical Record Text
• The Detection of Distributional Discrepancy for Text Generation
• Exposure Bias versus Self-Recovery: Are Distortions Really Incremental for Autoregressive Text Generation?
• Top-Down Tree Structured Text Generation