NCES

Implementation of neural class expression synthesizers (NCES)

Installation

Clone this repository:

https://github.com/ConceptLengthLearner/NCES.git

First install Anaconda3, then all required librairies by running the following:

conda env create -f environment.yml

A conda environment (cel) will be created. Next activate the environment: conda activate cel

Dowload DL-Learner-1.4.0 from github and extract it into the directory containing NCES (cloned above), not inside NCES!

Download Datasets from drive, extract it into NCES/Method and rename the folder as Datasets

Reproducing the reported results

NCES (Ours)

Open a terminal and navigate into Method/reproduce_results/ cd NCES/Method/reproduce_results/

• Reproduce training NCES: python reproduce_training_concept_synthesizers_[name_of_knowledge_graph].py

• Reproduce training NCES on all KGs: sh reproduce_training_nces_on_all_kgs.sh

• To reproduce evaluation results, please open the jupyter notebook/lab file ReproduceNCES.ipynb

DL-Learner

Open a terminal and navigate into Method/dllearner/ cd NCES/Method/dllearner/

• Reproduce CELOE, OCEL, and ELTL concept learning results: python reproduce_dllearner_experiment_[name_of_knowledge_graph].py

• Reproduce CELOE, OCEL, and ELTL results for all KGs: sh reproduce_dllearner_experiment_all_kgs.sh

Remark name_of_knowledge_graph is one of carcinogenesis_kg, semantic_bible_kg, mutagenesis_kg or family_benchmark_kg

Acknowledgement

We based our implementation on the open source implementation of ontolearn. We would like to thank the Ontolearn team for the readable codebase.