Paraphrase generation is an important problem in NLP, especially in question answering, information retrieval, information extraction, conversation systems, to name a few. In this paper, we address the problem of generating paraphrases automatically. Our proposed method is based on a combination of deep generative models (VAE) with sequence-to-sequence models (LSTM) to generate paraphrases, given an input sentence. Traditional VAEs when combined with recurrent neural networks can generate free text but they are not suitable for paraphrase generation for a given sentence. We address this problem by conditioning the both, encoder and decoder sides of VAE, on the original sentence, so that it can generate the given sentence's paraphrases. Unlike most existing models, our model is simple, modular and can generate multiple paraphrases, for a given sentence. Quantitative evaluation of the proposed method on a benchmark paraphrase dataset demonstrates its efficacy, and its performance improvement over the state-of-the-art methods by a significant margin, whereas qualitative human evaluation indicate that the generated paraphrases are well-formed, grammatically correct, and are relevant to the input sentence. Furthermore, we evaluate our method on a newly released question paraphrase dataset, and establish a new baseline for future research.
ParaSCI: A Large Scientific Paraphrase Dataset for Longer Paraphrase Generation
Wangxuan Institute of Computer Technology, Peking University(北京大学王选计算机研究所);The MOE Key Laboratory of Computational Linguistics, Peking University(北京大学计算语言学教育部重点实验室);College of Science, Minzu University of China(中央民族大学理学院);Center for Data Science, Peking University(北京大学数据科学中心)
Paraphrase Generation as Zero-Shot Multilingual Translation: Disentangling Semantic Similarity from Lexical and Syntactic Diversity
Johns Hopkins University(约翰斯·霍普金斯大学)
最近的研究表明,多语言神经机器翻译(NMT)模型可以用来判断一个句子对同一语言中另一个句子的释义程度(Thompson and Post,2020);然而,试图使用标准束搜索从这样的模型中生成释义会产生微不足道的副本或近似副本。我们引入了一个简单的释义生成算法,该算法不鼓励在输入中产生n元语法。我们的方法可以从一个单一的多语言NMT模型中生成多种语言的释义。此外,输入和输出之间的词汇多样性的量可以在生成时控制。我们进行了一次人工评估,将我们的方法与在大型英语合成释义数据库ParaBank 2上训练的释义器进行比较(Hu等人,2019 c),并发现我们的方法产生的释义,更好地保留意义,更符合语法,为同一水平的词汇多样性。其他较小的人类评估表明,我们的方法也适用于两种非英语语言。
ConRPG: Paraphrase Generation using Contexts as Regularizer
Key Lab of Intelligent Information Processing of Chinese Academy of Sciences(中国科学院智能信息处理重点实验室);Zhejiang University(浙江大学);Computer Center of Peking University(北京大学计算中心);Peng Cheng Laboratory(鹏城实验室)
Improved Paraphrase Generation via Controllable Latent Diffusion
National Key Laboratory for Novel Software Technology(计算机软件新技术国家重点实验室);Collaborative Innovation Center of Novel Software Technology and Industrialization(软件新技术与产业化协同创新中心);Nanjing University(南京大学)
Paraphrase Generation as Unsupervised Machine Translation
Zhejiang University(浙江大学);University of California, Los Angeles(加利福尼亚大学洛杉矶分校);Peng Cheng Laboratory(鹏城实验室);National Biomedical Imaging Center, Peking University(北京大学国家生物医学成像中心);Computer Center, Peking University(北京大学计算中心);Shanghai AI Laboratory(上海人工智能实验室);Shanghai Institute for Advanced Study of Zhejiang University(浙江大学上海高等研究院)
A task in a suit and a tie: paraphrase generation with semantic augmentation
The University of Texas at Austin(德克萨斯大学奥斯汀分校);Google AI Language(谷歌人工智能语言团队)
释义的基础是语义。我们展示了Transformers(Vaswani et al. 2017)用于释义生成的有效性,并通过多编码器合并PropBank标签进行了进一步改进。在MSCOCO和WikiAnswers上进行评估,我们发现Transformers是快速有效的,并且Transformers和LSTM的语义增强导致BLEU,METEOR和TER的2-3点增益。更重要的是,与以前的模型相比,我们发现人类评估的收益惊人地大。尽管如此,对生成的释义进行手动检查,发现有很大的改进空间:即使是我们最好的模型,也只能为CHIA数据集中28%的标题生成人类可接受的释义(Sharma et al. 2018),而且它在维基百科的句子上也失败了。总的来说,这些结果表明,在任务中纳入语义的潜力,同时强调需要更强的评估。
Adversarial Example Generation with Syntactically Controlled Paraphrase Networks
Allen Institute of Artificial Intelligence(艾伦人工智能研究所);UMass Amherst(马萨诸塞大学阿默斯特分校);Carnegie Mellon University(卡内基梅隆大学);Toyota Technological Institute at Chicago(芝加哥丰田理工学院);University of Washington(华盛顿大学)
Pushing Paraphrase Away from Original Sentence: A Multi-Round Paraphrase Generation Approach
Wangxuan Institute of Computer Technology, Peking University(北京大学王选计算机研究所);Center for Data Science, Peking University(北京大学数据科学研究中心);The MOE Key Laboratory of Computational Linguistics, Peking University(北京大学计算语言学教育部重点实验室)
ParaAMR: A Large-Scale Syntactically Diverse Paraphrase Dataset by AMR Back-Translation
University of California Los Angeles(加州大学洛杉矶分校);University of Illinois Chicago(伊利诺伊大学芝加哥分校);University of Southern California(南加州大学);Amazon Alexa AI(亚马逊Alexa人工智能部门)
Towards Document-Level Paraphrase Generation with Sentence Rewriting and Reordering
Wangxuan Institute of Computer Technology, Peking University(北京大学王选计算机研究所);Center for Data Science, Peking University(北京大学数据科学研究中心);The MOE Key Laboratory of Computational Linguistics, Peking University(北京大学计算语言学教育部重点实验室)
Visual Information Guided Zero-Shot Paraphrase Generation
Wangxuan Institute of Computer Technology, Peking University(北京大学王选计算机研究所);Center for Data Science, Peking University(北京大学数据科学中心);The MOE Key Laboratory of Computational Linguistics, Peking University(北京大学计算语言学教育部重点实验室)
Tsinghua University(清华大学);University of California, Davis(加州大学戴维斯分校)
如果给定一个巨大的数据集,神经生成模型在对话生成任务上已经取得了很好的性能。然而,缺乏高质量的对话框数据和昂贵的数据注释过程极大地限制了它们在现实世界中的应用。我们提出了一个释义增强响应生成(PARG)框架,联合训练的释义模型和响应生成模型,以提高对话生成性能。设计了一种基于对话状态和对话行为标签的释义训练数据自动构造方法。PARG适用于各种对话生成模型,诸如TSCP(Lei等人,2018)和DAMD(Zhang et al.,2019年)。实验结果表明,该框架在CamRest676和MultiWOZ上进一步改进了这些最先进的对话模型。PARG在对话框生成任务中的表现也明显优于其他数据增强方法,特别是在低资源设置下。
Towards Diverse Paraphrase Generation Using Multi-Class Wasserstein GAN
Tsinghua University(清华大学);Department of Electronic Engineering, Tsinghua University(清华大学电子工程系);Xiamen University(厦门大学);Department of Communication Engineering, Xiamen University(厦门大学通信工程系);Key Laboratory of Digital Fujian on IoT Communication, Architecture and Security Technology, Xiamen University(厦门大学福建省数字物联网通信、架构与安全技术重点实验室)