Unsupervised Paraphrase Generation using Pre-trained Language Models
arXiv:2006.05477
Abstract
Large scale Pre-trained Language Models have proven to be very powerful approach in various Natural language tasks. OpenAI's GPT-2 \cite{radford2019language} is notable for its capability to generate fluent, well formulated, grammatically consistent text and for phrase completions. In this paper we leverage this generation capability of GPT-2 to generate paraphrases without any supervision from labelled data. We examine how the results compare with other supervised and unsupervised approaches and the effect of using paraphrases for data augmentation on downstream tasks such as classification. Our experiments show that paraphrases generated with our model are of good quality, are diverse and improves the downstream task performance when used for data augmentation.
References in corpus (8)
- HuggingFace's Transformers: State-of-the-art Natural Language Processing
- Neural Discrete Representation Learning
- Neural Paraphrase Generation with Stacked Residual LSTM Networks
- Hierarchical Neural Story Generation
- A Deep Generative Framework for Paraphrase Generation
- Joint Copying and Restricted Generation for Paraphrase
- Learning to Paraphrase for Question Answering
- Unsupervised Paraphrasing by Simulated Annealing
Cited by in corpus (6)
- LM-CPPF: Paraphrasing-Guided Data Augmentation for Contrastive Prompt-Based Few-Shot Fine-Tuning
- Automated Annotation of Scientific Texts for ML-based Keyphrase Extraction and Validation
- Paraphrase Generation as Unsupervised Machine Translation
- SMCLM: Semantically Meaningful Causal Language Modeling for Autoregressive Paraphrase Generation
- ConRPG: Paraphrase Generation using Contexts as Regularizer
- Unsupervised Paraphrasing with Pretrained Language Models