Integrating Relation Constraints with Neural Relation Extractors
Authors: Yuan Ye, Yansong Feng, Bingfeng Luo, Yuxuan Lai, Dongyan Zhao9442-9449
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on both English and Chinese datasets show that our approach can help NNs learn from discrete relation constraints to reduce inconsistency among local predictions, and outperform popular neural relation extraction (NRE) models even enhanced with extra post-processing. |
| Researcher Affiliation | Academia | Yuan Ye, Yansong Feng,* Bingfeng Luo, Yuxuan Lai, Dongyan Zhao Wangxuan Institute of Computer Technology, Peking University, China {pkuyeyuan, fengyansong, bf luo, erutan, zhaodongyan}@pku.edu.cn |
| Pseudocode | No | The paper describes the methods 'Coherent' and 'Semantic' using mathematical equations and textual descriptions, but it does not include an explicitly labeled 'Pseudocode' or 'Algorithm' block. |
| Open Source Code | Yes | Our source code and datasets will be released at https://github.com/PKUYe Yuan/ Constraint-Loss-AAAI-2020. |
| Open Datasets | Yes | We evaluate our approach on both English and Chinese datasets constructed by Chen et al. 2018. The English one is constructed by mapping triples in DBpedia (Bizer et al. 2009) to sentences in the New York Times Corpus. It has 51 relations, about 50k triples, 134k sentences for training and 30k triples, 53k sentences for testing. The Chinese dataset is built by mapping the triples of Hudong Bai Ke, a large Chinese encyclopedia, with four Chinese economic newspapers. It contains 28 relations, about 60k triples, 120k sentences for training and 40k triples, 83k sentences for testing. Our source code and datasets will be released at https://github.com/PKUYe Yuan/ Constraint-Loss-AAAI-2020. |
| Dataset Splits | No | The English one ... has 50k triples, 134k sentences for training and 30k triples, 53k sentences for testing. The Chinese dataset ... contains 28 relations, about 60k triples, 120k sentences for training and 40k triples, 83k sentences for testing. The paper clearly states training and testing set sizes but does not mention a validation set split. |
| Hardware Specification | No | The paper does not provide any specific hardware details such as GPU models, CPU types, or memory specifications used for running the experiments. |
| Software Dependencies | No | The paper discusses neural network models and architectures like CNN and PCNN, but it does not specify any software dependencies with version numbers, such as programming languages, libraries, or frameworks (e.g., Python, PyTorch, TensorFlow versions). |
| Experiment Setup | No | The paper states 'We use a grid search to tune our hyper parameters, including the weight coefficient λ. Details about our hyper parameters are reported in Appendix.' This indicates that the specific details are in an appendix not provided, rather than in the main body of the text. |