Simple and Effective Relation-based Embedding Propagation for Knowledge Representation Learning

Authors: Huijuan Wang, Siming Dai, Weiyue Su, Hui Zhong, Zeyang Fang, Zhengjie Huang, Shikun Feng, Zeyu Chen, Yu Sun, Dianhai Yu

IJCAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments also demonstrate that REP has significant scalability while improving or maintaining prediction quality. Notably, it averagely brings about 10% relative improvement to triplet-based embedding methods on OGBLWiki KG2 and takes 5%-83% time to achieve comparable results as the state-of-the-art GC-OTE.
Researcher Affiliation Industry Huijuan Wang , Siming Dai , Weiyue Su , Hui Zhong , Zeyang Fang , Zhengjie Huang , Shikun Feng , Zeyu Chen , Yu Sun and Dianhai Yu Baidu, Inc. {wanghuijuan03, daisiming, suweiyue, zhonghui03, fangzeyang, huangzhengjie, fengshikun01, chenzeyu01, sunyu02, yudianhai}@baidu.com
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code Yes The code and details of REP are released in Graph4KG 3. https://github.com/PaddlePaddle/PGL/tree/main/apps/Graph4KG
Open Datasets Yes OGBL-Wiki KG2 [Hu et al., 2020] is a link prediction dataset from Open Graph Benchmark (OGB) 2. Wiki KG90M-LSC [Hu et al., 2021] of KDD Cup 2021 1 is extracted from Wikidata.
Dataset Splits Yes Wiki KG90M-LSC... provides 1000 randomly sampled entity candidates for each tail in the validation/test set. OGBL-Wiki KG2... 1000 randomly sampled candidates for tail entities are provided likewise.
Hardware Specification Yes Experiments were conducted on Intel Xeon Gold 6271C CPUs and Tesla V100 SXM2 GPUs.
Software Dependencies No The paper mentions using "public code to reproduce triplet-based methods" and their own code on "PaddlePaddle/PGL" but does not specify software dependencies with version numbers.
Experiment Setup Yes To investigate the impact of α, we set α as values ranging from 0.95 to 0.99 with step 0.01. We investigate how multi-hop neighbors influence embedding quality by changing the number of hops from 1 to 20. We set α = 0.98 to control variables.