Pykg2vec: A Python Library for Knowledge Graph Embedding
Authors: Shih-Yuan Yu, Sujit Rokka Chhetri, Arquimedes Canedo, Palash Goyal, Mohammad Abdullah Al Faruque
JMLR 2021 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Pykg2vec’s flexible and modular software architecture currently implements 25 state-of-the-art knowledge graph embedding algorithms... The results of running the mentioned scripts are shown in Table 2 and Table 3. Table 2 demonstrates the performance of KGE methods, while Table 3 shows the effect of utilizing the setting found by Bayesian Optimizer and the comparison with other KGE libraries. |
| Researcher Affiliation | Collaboration | Shih-Yuan Yu EMAIL Sujit Rokka Chhetri EMAIL Department of Electrical Engineering and Computer Science, University of California-Irvine Arquimedes Canedo EMAIL Siemens Corporate Technology, Princeton Palash Goyal EMAIL Department of Computer Science, University of Southern California Mohammad Abdullah Al Faruque EMAIL Department of Electrical Engineering and Computer Science, University of California-Irvine |
| Pseudocode | No | The paper discusses the software architecture of pykg2vec (Figure 1) and provides usage examples as command-line snippets. However, it does not include any clearly labeled pseudocode or algorithm blocks describing the underlying methods. |
| Open Source Code | Yes | The source code of pykg2vec is available at https://github.com/Sujit-O/pykg2vec . |
| Open Datasets | Yes | The results of KGE methods on FB15k using the settings in the original papers. The effect of applying the found hyperparameter setting on Trans E for WN18 RR. |
| Dataset Splits | No | The KG Controller module handles low-level parsing tasks such as finding the total unique set of entities and relations; creating ordinal encoding maps; performing train-test split; and caching the data on disk to optimize tasks that involve repetitive model testing. This mentions "train-test split" but gives no specific details about the methodology (e.g., percentages, random seed, or file names). |
| Hardware Specification | No | Pykg2vec is built with Python and Py Torch that allows the computations to be assigned on GPUs (legacy Tensor Flow version is also ready in a separate branch). This mentions "GPUs" but does not specify any particular model or other hardware specifications. |
| Software Dependencies | No | Pykg2vec is built on top of Py Torch and Python’s multiprocessing framework and provides modules for batch generation, Bayesian hyperparameter optimization, evaluation of KGE tasks, embedding, and result visualization. This mentions software like Python and Py Torch but does not specify their version numbers. |
| Experiment Setup | Yes | Found Golden Setting : # a f t e r at max 100 t r i a l s . { L1 flag : True , batch size : 2279 , hidden size : 80 , l e a r n i n g r a t e : 0.05314 , margin : 8.58 , opt : adam } Table 2 demonstrates the performance of KGE methods on FB15k using the settings in the original papers. |