Type-augmented Relation Prediction in Knowledge Graphs

Authors: Zijun Cui, Pavan Kapanipathi, Kartik Talamadupula, Tian Gao, Qiang Ji7151-7159

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our proposed Ta RP method achieves significantly better performance than state-of-the-art methods on four benchmark datasets: FB15K, FB15K-237, YAGO26K-906, and DB111K-174.
Researcher Affiliation Collaboration Zijun Cui,1 Pavan Kapanipathi,2 Kartik Talamadupula,2 Tian Gao,2 Qiang Ji1 1Rensselaer Polytechnic Institute 2 IBM Research
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide concrete access to source code for the methodology described in this paper.
Open Datasets Yes We consider three benchmark datasets for the relation prediction task: FB15K (Bordes et al. 2013), YAGO26K-906 (Hao et al. 2019) and DB111K-74 (Hao et al. 2019).
Dataset Splits Yes On each dataset, we select the threshold η from {0, 0.1, 0.2, 0.4, 0.6, 0.8, 0.9} that achieves the best relation prediction performance (Hits@1) on the validation set.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers like Python 3.8, CPLEX 12.4) needed to replicate the experiment.
Experiment Setup Yes Ta RP has one hyper-parameter threshold η. On each dataset, we select the threshold η from {0, 0.1, 0.2, 0.4, 0.6, 0.8, 0.9} that achieves the best relation prediction performance (Hits@1) on the validation set. On FB15K and YAGO26K-906, η = 0.1. On DB111K-174, η = 0.