Learning Paraphrase Identification with Structural Alignment
Authors: Chen Liang, Praveen Paritosh, Vinodh Rajendran, Kenneth D. Forbus
IJCAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our approach is evaluated on the paraphrase identification task and achieved results competitive with the state-of-the-art. |
| Researcher Affiliation | Collaboration | Chen Liang,1 Praveen Paritosh,2 Vinodh Rajendran,2 Kenneth D. Forbus1 1Northwestern University, Evanston, IL 2Google Research, Mountain View, CA |
| Pseudocode | No | The paper describes its methods using prose and mathematical equations but does not include any pseudocode or clearly labeled algorithm blocks. |
| Open Source Code | No | The paper mentions 'pretrained Word2Vec word embeddings1 [Mikolov et al., 2013]' and provides a link in footnote 1: 'https://code.google.com/archive/p/word2vec/'. This link refers to a third-party tool used by the authors, not the open-source code for their own described methodology. |
| Open Datasets | Yes | The dataset is MSRP paraphrase corpus [Dolan and Brockett, 2005]. |
| Dataset Splits | Yes | We used the training and test split provided by the corpus (4076 training and 1725 test examples), and selected 20% of the training data by stratified sampling as the validation set. |
| Hardware Specification | No | The paper does not specify any hardware details such as CPU/GPU models, memory, or cloud computing instances used for running the experiments. |
| Software Dependencies | No | The paper mentions using 'Stanford Core NLP' and 'pretrained Word2Vec word embeddings' but does not provide specific version numbers for these or any other software components, which is necessary for reproducibility. |
| Experiment Setup | Yes | We run averaged structured perceptron for 10 epochs. The averaged parameters after each epoch is stored as {wi, i = 1, 2, ..., 10}, and we used a validation set to decide which one to use as the final parameters. We also used the validation set to decide when to stop the iterative training. |