Controllable Natural Language Generation with Contrastive Prefixes
TLDR
A novel lightweight framework for controllable GPT2 generation, which utilizes a set of small attribute-specific vectors, called prefixes, to steer natural language generation, and shows that its methods can guide generation towards the desired attributes while keeping high linguistic quality.Abstract:
To guide the generation of large pretrained language models (LM), previous work has focused on directly fine-tuning the language model or utilizing an attribute discriminator. In this work, we propose a novel lightweight framework for controllable GPT2 generation, which utilizes a set of small attribute-specific vectors, called prefixes (Li and Liang, 2021), to steer natural language generation. Different from Li and Liang (2021), where each prefix is trained independently, we take the relationship among prefixes into consideration and train multiple prefixes simultaneously. We propose a novel supervised method and also an unsupervised method to train the prefixes for single-aspect control while the combination of these two methods can achieve multi-aspect control. Experimental results on both single-aspect and multi-aspect control show that our methods can guide generation towards the desired attributes while keeping high linguistic quality.read more
Citations
More filters
Proceedings ArticleDOI
RLPrompt: Optimizing Discrete Text Prompts with Reinforcement Learning
Mingkai Deng,Jianyu Wang,Cheng-Ping Hsieh,Han Guo,Tianmin Shu,Meng Song,Eric P. Xing,Zhiting Hu +7 more
TL;DR: RLP ROMPT formulates a parameter-efficient policy network that generates the desired discrete prompt after training with reward to overcome the complexity and stochasticity of reward signals by the large LM environment, and incorporates effective reward stabilization that substantially enhances the training efflciency.
Journal ArticleDOI
Tailor: A Prompt-Based Approach to Attribute-Based Controlled Text Generation
TL;DR: The proposed Tailor represents each attribute as a pre-trained continuous vector and guides the generation of a PLM switch to aPre-specified attribute and introduces a trainable prompt connector, which can be concatenated with any two single-attribute prompts to multi-attribute text generation.
Proceedings ArticleDOI
Generative Knowledge Graph Construction: A Review
TL;DR: This study summarizes the recent compelling progress in generative knowledge graph construction and presents a detailed, complete taxonomy for the generative KGC methods.
Proceedings Article
Gradient-based Constrained Sampling from Language Models
TL;DR: This article proposed a sampling procedure that combines the log-likelihood of the language model with arbitrary (differentiable) constraints in a single energy function, and then generates samples in a non-autoregressive manner.
Proceedings ArticleDOI
A Distributional Lens for Multi-Aspect Controllable Text Generation
TL;DR: Experiments on the three-aspect control task reveal that the proposed directly search for the intersection areas of multiple attribute distributions as their combination for generation outperforms several strong baselines on attribute relevance and text quality and achieves the SOTA.
References
More filters
Posted Content
RoBERTa: A Robustly Optimized BERT Pretraining Approach
Yinhan Liu,Myle Ott,Naman Goyal,Jingfei Du,Mandar Joshi,Danqi Chen,Omer Levy,Michael Lewis,Luke Zettlemoyer,Veselin Stoyanov +9 more
TL;DR: It is found that BERT was significantly undertrained, and can match or exceed the performance of every model published after it, and the best model achieves state-of-the-art results on GLUE, RACE and SQuAD.
Proceedings Article
Learning Word Vectors for Sentiment Analysis
TL;DR: This work presents a model that uses a mix of unsupervised and supervised techniques to learn word vectors capturing semantic term--document information as well as rich sentiment content, and finds it out-performs several previously introduced methods for sentiment classification.
Proceedings Article
Categorical Reparameterization with Gumbel-Softmax
Eric Jang,Shixiang Gu,Ben Poole +2 more
TL;DR: Gumbel-Softmax as mentioned in this paper replaces the non-differentiable samples from a categorical distribution with a differentiable sample from a novel Gumbel softmax distribution, which has the essential property that it can be smoothly annealed into the categorical distributions.
Proceedings Article
Neural Discrete Representation Learning
TL;DR: The Vector Quantised-Variational AutoEncoder (VQ-VAE) as discussed by the authors is a generative model that learns a discrete latent representation by using vector quantization.
Posted Content
Character-level Convolutional Networks for Text Classification
TL;DR: This article constructed several large-scale datasets to show that character-level convolutional networks could achieve state-of-the-art or competitive results in text classification.
Related Papers (5)
An Investigation of Fine-tuning Pre-trained Model for MR-to-Text Generation
Ting Hu,Christoph Meinel +1 more