PyKEEN

PyKEEN (Python KnowlEdge EmbeddiNgs) is a Python package designed to train and evaluate knowledge graph embedding models (incorporating multi-modal information). It is part of the KEEN Universe.

Installation • Quickstart • Models • Data Sets • Support

Installation

The development version of PyKEEN (POEM) can be downloaded and installed from GitHub on Python 3.7+ with:

$ git clone https://github.com/mali-git/POEM_develop.git poem
$ cd poem
$ pip install -e .

Quickstart

This example shows how to train a model on a data set and test on another data set.

The fastest way to get up and running is to use the pipeline function. It provides a high-level entry into the extensible functionality of this package. The following example shows how to train and evaluate the TransE model on the Nations dataset. By default, the training loop uses the open world assumption and evaluates with rank-based evaluation.

from poem.pipeline import pipeline
result = pipeline(
    model='TransE',
    data_set='nations',
)

The results are returned in a dataclass that has attributes for the trained model, the training loop, and the evaluation.

POEM is extensible such that:

• Each model has the same API, so anything from poem.models can be dropped in
• Each training loop has the same API, so poem.training.CWATrainingLoop can be dropped in
• Triples factories can be generated by the user with from poem.triples.TriplesFactory

Implementation

Models

	Name	Reference	Citation
1	ComplEx	poem.models.ComplEx	Trouillon et al., 2016
2	ComplExLiteralCWA	poem.models.ComplExLiteralCWA	Agustinus et al., 2018
3	ConvE	poem.models.ConvE	Dettmers et al., 2018
4	ConvKB	poem.models.ConvKB	Nguyen et al., 2018
5	DistMult	poem.models.DistMult	Yang et al., 2014
6	DistMultLiteral	poem.models.DistMultLiteral	Agustinus et al., 2018
7	ERMLP	poem.models.ERMLP	Dong et al., 2014
8	HolE	poem.models.HolE	Nickel et al., 2016
9	KG2E	poem.models.KG2E	He et al., 2015
10	NTN	poem.models.NTN	Socher et al., 2013
11	ProjE	poem.models.ProjE	Shi et al., 2017
12	RESCAL	poem.models.RESCAL	Nickel et al., 2011
13	RotatE	poem.models.RotatE	Sun et al., 2019
14	SimplE	poem.models.SimplE	Kazemi et al., 2018
15	StructuredEmbedding	poem.models.StructuredEmbedding	Bordes et al., 2011
16	TransD	poem.models.TransD	Ji et al., 2015
17	TransE	poem.models.TransE	Bordes et al., 2013
18	TransH	poem.models.TransH	Wang et al., 2014
19	TransR	poem.models.TransR	Lin et al., 2015
20	TuckEr	poem.models.TuckEr	Balazevic et al., 2019
21	UnstructuredModel	poem.models.UnstructuredModel	Bordes et al., 2014

Data Sets

	Name	Reference	Description
1	fb15k	poem.datasets.fb15k	The FB15k data set.
2	fb15k237	poem.datasets.fb15k237	The FB15k-237 data set.
3	kinship	poem.datasets.kinship	The Kinship data set.
4	nations	poem.datasets.nations	The Nations data set.
5	umls	poem.datasets.umls	The UMLS data set.
6	wn18	poem.datasets.wn18	The WN18 data set.
7	wn18rr	poem.datasets.wn18rr	The WN18-RR data set.
8	yago310	poem.datasets.yago310	The YAGO3-10 data set is a subset of YAGO3 that only contains entities with at least 10 relations.

Training Modes

	Name	Reference	Description
1	cwa	poem.training.CWATrainingLoop	A training loop that uses the closed world assumption.
2	owa	poem.training.OWATrainingLoop	A training loop that uses the open world assumption.

Evaluators

	Name	Reference	Description
1	rankbased	poem.evaluators.RankBasedEvaluator	A rank-based evaluator for KGE models.

Metrics

	Name	Reference	Description
1	metric	poem.evaluators.MetricResults	Results from computing metrics.

Supporters

This project has been supported by several organizations:

• Smart Data Analytics (University of Bonn)
• Fraunhofer Institute for Intelligent Analysis and Information Systems
• Bonn Aachen International Center for IT (University of Bonn)
• Fraunhofer Institute for Algorithms and Scientific Computing
• Fraunhofer Center for Machine Learning
• Technical University of Denmark - DTU Compute - Section for Cognitive Systems
• Technical University of Denmark - DTU Compute - Section for Statistics and Data Analysis