Inductive Relation Prediction by Subgraph Reasoning

Authors: Komal Teru, Etienne Denis, Will Hamilton

ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We perform experiments on three benchmark knowledge completion datasets: WN18RR (Dettmers et al., 2018), FB15k-237 (Toutanova et al., 2015), and NELL-995 (Xiong et al., 2017) (and other variants derived from them). Our empirical study is motivated by the following questions: 1. Inductive relation prediction. (...) 2. Transductive relation prediction. (...) 3. Ablation study.
Researcher Affiliation Academia Komal K. Teru 1 2 Etienne G. Denis 1 2 William L. Hamilton 1 2 1Mc Gill University 2Mila. Correspondence to: Komal K. Teru <komal.teru@mail.mcgill.ca>.
Pseudocode No The paper describes the model details and training regime using descriptive text and mathematical equations, but does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code Yes The code and the data for all the following experiments is available at: https://github.com/kkteru/grail.
Open Datasets Yes We perform experiments on three benchmark knowledge completion datasets: WN18RR (Dettmers et al., 2018), FB15k-237 (Toutanova et al., 2015), and NELL-995 (Xiong et al., 2017) (and other variants derived from them).
Dataset Splits Yes For NELL-995, we split the whole dataset into train/valid/test set by the ratio 70/15/15, making sure all the entities and relations in the valid and test splits occur at least once in the train set.
Hardware Specification No The paper does not specify the hardware used for experiments, such as specific GPU or CPU models, memory, or cloud computing instances with detailed specifications.
Software Dependencies No The paper does not provide specific software dependencies or their version numbers, such as programming language versions or library versions (e.g., Python 3.x, PyTorch 1.x, CUDA 11.x).
Experiment Setup Yes We employ a 3-layer GNN with the dimension of all latent embeddings equal to 32. The basis dimension is set to 4 and the edge dropout rate to 0.5.