MARGE: Pre-training via Paraphrasing
A retrieval model maps a document to a set of related documents, which a reconstruction model paraphrases to maximize the likelihood of the original.
paraphrasing pretraining marge natural-language-processing research paper arxiv:2006.15020

We introduce MARGE, a pre-trained sequence-to-sequence model learned with an unsupervised multi-lingual multi-document paraphrasing objective. MARGE provides an alternative to the dominant masked language modeling paradigm, where we self-supervise the reconstruction of target text by retrieving a set of related texts (in many languages) and conditioning on them to maximize the likelihood of generating the original. We show it is possible to jointly learn to do retrieval and reconstruction, given only a random initialization. The objective noisily captures aspects of paraphrase, translation, multi-document summarization, and information retrieval, allowing for strong zero-shot performance on several tasks. For example, with no additional task-specific training we achieve BLEU scores of up to 35.8 for document translation. We further show that fine-tuning gives strong performance on a range of discriminative and generative tasks in many languages, making MARGE the most generally applicable pre-training method to date.

Don't forget to tag @sidaw in your comment.

Authors community post
Share this project
Similar projects
Paraphrase Any Question with T5 (Text-To-Text Transformer)
Given a question, generate paraphrased versions of the question with T5 transformer. Pretrained model and training script provided.