Dictionary-Guided Editing Networks for Paraphrase Generation
Authors: Shaohan Huang, Yu Wu, Furu Wei, Zhongzhi Luan6546-6553
AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct experiments on two benchmark datasets for paraphrase generation, namely the MSCOCO and Quora dataset. The automatic evaluation results demonstrate that our dictionary-guided editing networks outperforms the baseline methods. |
| Researcher Affiliation | Collaboration | Shaohan Huang, Yu Wu, Furu Wei, Zhongzhi Luan Sino-German Joint Software Institute, Beihang University, Beijing, China Microsoft Research, Beijing, China State Key Lab of Software Development Environment, Beihang University, Beijing, China {shaohanh, fuwei}@microsoft.com wuyu@buaa.edu.cn |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks explicitly labeled as 'Pseudocode' or 'Algorithm'. |
| Open Source Code | No | The paper mentions a link to the code for a baseline model ('Residual LSTM') but does not provide concrete access to the source code for their own proposed methodology. |
| Open Datasets | Yes | MSCOCO (Lin et al. 2014) is a large-scale captioning dataset which contains human annotated captions of over 120K images 2. This dataset was used previously to evaluate paraphrase generation methods (Prakash et al. 2016; Gupta et al. 2018). ... Quora dataset is related to the problem of identifying duplicate questions3. It consists of over 400K potential question duplicate pairs. ... 2http://cocodataset.org/ 3https://data.quora.com/First-Quora-Dataset-Release Question-Pairs |
| Dataset Splits | Yes | For the MSCOCO dataset: 20K instances are randomly selected from the data for testing, 10K instances for validation and remaining data over 320K instances for training. For the Quora dataset: 140K instances are randomly selected for training, 5K instances for validation and about 5K instances for testing. |
| Hardware Specification | No | The paper does not provide specific hardware details (such as exact GPU/CPU models, processor types, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper mentions software like NLTK and Open NMT but does not provide specific version numbers for these or any other ancillary software components required for reproducibility. |
| Experiment Setup | Yes | The dimensions of word embeddings is set to 300 and hidden vectors are set to 512 in the sequence encoder and decoder. The dimensions of the attention vector are also set to 512 and the dropout rate is set to 0.5 for regularization. The mini-batched Adam (Kingma and Ba 2014) algorithm is used to optimize the objective function. The batch size and base learning rates are set to 64 and 0.0001, respectively. |