{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/a-deep-generative-framework-for-paraphrase","title":"A Deep Generative Framework for Paraphrase Generation","arxiv_id":"1709.05074","date":"2017-09-15","proceeding":null,"authors":["Ankush Gupta","Arvind Agarwal","Prawaan Singh","Piyush Rai"],"abstract":"Paraphrase generation is an important problem in NLP, especially in question\nanswering, information retrieval, information extraction, conversation systems,\nto name a few. In this paper, we address the problem of generating paraphrases\nautomatically. Our proposed method is based on a combination of deep generative\nmodels (VAE) with sequence-to-sequence models (LSTM) to generate paraphrases,\ngiven an input sentence. Traditional VAEs when combined with recurrent neural\nnetworks can generate free text but they are not suitable for paraphrase\ngeneration for a given sentence. We address this problem by conditioning the\nboth, encoder and decoder sides of VAE, on the original sentence, so that it\ncan generate the given sentence's paraphrases. Unlike most existing models, our\nmodel is simple, modular and can generate multiple paraphrases, for a given\nsentence. Quantitative evaluation of the proposed method on a benchmark\nparaphrase dataset demonstrates its efficacy, and its performance improvement\nover the state-of-the-art methods by a significant margin, whereas qualitative\nhuman evaluation indicate that the generated paraphrases are well-formed,\ngrammatically correct, and are relevant to the input sentence. Furthermore, we\nevaluate our method on a newly released question paraphrase dataset, and\nestablish a new baseline for future research.","url_abs":"http://arxiv.org/abs/1709.05074v1","url_pdf":"http://arxiv.org/pdf/1709.05074v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"a-deep-generative-framework-for-paraphrase","repo_url":"https://github.com/arvind385801/paraphrasegen","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"information-retrieval","task_name":"Information Retrieval"},{"task_slug":"paraphrase-generation","task_name":"Paraphrase Generation"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":"sentence","task_name":"Sentence"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1709.05074","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}