SimplE Embedding for Link Prediction in Knowledge Graphs
Authors: Seyed Mehran Kazemi, David Poole
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We show empirically that, despite its simplicity, Simpl E outperforms several state-of-the-art tensor factorization techniques. Simpl E s code is available on Git Hub at https://github.com/Mehran-k/Simpl E. We conducted experiments on two standard benchmarks: WN18 a subset of Wordnet [24], and FB15k a subset of Freebase [2]. Table 1 shows the results of our experiments. |
| Researcher Affiliation | Academia | Seyed Mehran Kazemi University of British Columbia Vancouver, BC, Canada smkazemi@cs.ubc.ca David Poole University of British Columbia Vancouver, BC, Canada poole@cs.ubc.ca |
| Pseudocode | No | The paper describes the model definition and learning process in text and mathematical formulas but does not provide pseudocode or a clearly labeled algorithm block. |
| Open Source Code | Yes | Simpl E s code is available on Git Hub at https://github.com/Mehran-k/Simpl E. |
| Open Datasets | Yes | Datasets: We conducted experiments on two standard benchmarks: WN18 a subset of Wordnet [24], and FB15k a subset of Freebase [2]. |
| Dataset Splits | Yes | WN18 contains 40, 943 entities, 18 relations, 141, 442 train, 5, 000 validation and 5, 000 test triples. FB15k contains 14, 951 entities, 1, 345 relations, 483, 142 train, 50, 000 validation, and 59, 071 test triples. |
| Hardware Specification | No | The paper mentions implementing Simpl E in Tensor Flow but does not specify any hardware details like GPU/CPU models, memory, or cloud resources used for running experiments. |
| Software Dependencies | No | We implemented Simpl E in Tensor Flow [1]. While TensorFlow is mentioned, no specific version number is provided for TensorFlow or any other software dependencies. |
| Experiment Setup | Yes | We used the same search grid on embedding size and λ as [39] to make our results directly comparable to their results. We fixed the maximum number of iterations to 1000 and the number of batches to 100. We set the learning rate for WN18 to 0.1 and for FB15k to 0.05 and used adagrad to update the learning rate after each batch. Following [39], we generated one negative example per positive example for WN18 and 10 negative examples per positive example in FB15k. The best embedding size and λ values on WN18 for Simpl E-ignr were 200 and 0.001 respectively, and for Simpl E were 200 and 0.03. The best embedding size and λ values on FB15k for Simpl E-ignr were 200 and 0.03 respectively, and for Simpl E were 200 and 0.1. |