End-to-End Structure-Aware Convolutional Networks for Knowledge Base Completion

Authors: Chao Shang, Yun Tang, Jing Huang, Jinbo Bi, Xiaodong He, Bowen Zhou3060-3067

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the effectiveness of the proposed SACN on standard FB15k-237 and WN18RR datasets, and show about 10% relative improvement over the stateof-the-art Conv E in terms of HITS@1, HITS@3 and HITS@10. Table 3: Link prediction for FB15k-237, WN18RR and FB15k-237-Attr datasets.
Researcher Affiliation Collaboration 1Department of Computer Science and Engineering, University of Connecticut, Storrs, CT, USA 2JD AI Research, Mountain View, CA, USA
Pseudocode No The paper does not contain any clearly labeled pseudocode or algorithm blocks.
Open Source Code Yes The code for our model and experiments is publicly available 1. 1https://github.com/JD-AI-Research-Silicon-Valley/SACN
Open Datasets Yes Three benchmark datasets (FB15k-237, WN18RR and FB15k-237-Attr) are utilized in this study to evaluate the performance of link prediction. FB15k-237. The FB15k-237 (Toutanova and Chen 2015) dataset... WN18RR. WN18RR (Dettmers et al. 2017) is created from WN18 (Bordes et al. 2013)... FB15k-237-Attr. We extract the attribute triples of entities in FB15k-237 from FB24k. FB24k (Lin, Liu, and Sun 2016) is built based on Freebase dataset.
Dataset Splits Yes Each dataset is split into three sets for: training, validation and testing, which is same with the setting of the original Conv E. Table 2: Statistics of datasets. Dataset... Train Edges... Val. Edges... Test Edges...
Hardware Specification Yes Our models are implemented by Py Torch and run on NVIDIA Tesla P40 Graphics Processing Units.
Software Dependencies No The paper mentions 'Py Torch' but does not provide specific version numbers for software dependencies.
Experiment Setup Yes The hyperparameters in our Conv-Trans E and SACN models are determined by a grid search during the training. We manually specify the hyperparameter ranges: learning rate {0.01, 0.005, 0.003, 0.001}, dropout rate {0.0, 0.1, 0.2, 0.3, 0.4, 0.5}, embedding size {100, 200, 300}, number of kernels {50, 100, 200, 300}, and kernel size {2 1, 2 3, 2 5}. Here all the models use the WGCN with two layers. For different datasets, we have found that the following settings work well: for FB15k-237, set the dropout to 0.2, number of kernels to 100, learning rate to 0.003 and embedding size to 200 for SACN; for WN18RR dataset, set dropout to 0.2, number of kernels to 300, learning rate to 0.003, and embedding size to 200 for SACN.