AMD-v2 OAEI 2022
AgreementMakerDeep (AMD) is a new flexible and extensible ontology matching system with PLM and KGE techniques. In OAEI 2022, we only apply zero-shot learning in Bio-ML track. AMD achieved competetive performance in terms of the evaluateion metrics used in the track without extra training. We also upload our alignments results of all dataset in Bio-ML track for reference.
- Python >= 3.7:
We recommend you to use Anaconda to create a conda environment: conda create -n AMD-Seals python=3.7
conda activate AMD-Seals
-
Other requirements:
pip install -r requirements.txt
Note: This file is able to produce alignments without MELT package.
-
To evaluate it with previous seals client and MELT Track repository:
java -jar seals-omt-client.jar AMD-seals -x http:https://oaei.webdatacommons.org/tdrs/ Suite-ID Version-ID /Users/AMD -a
examples: java -jar /Users/Ellen/Downloads/seals-omt-client.jar /Users/Ellen/Desktop/AMDSeals/target/AMD-seals -x http:https://oaei.webdatacommons.org/tdrs/ largebio largebio-snomed_nci_small_2016 /Users/Ellen/Downloads/AMD -a
-
Re-usage
If any changes to the model, then use MELT to wrapped and evalute.
For the whole AMD-seals, please refer to AMD for OAEI 2021.
If you find this repo useful and use our code for your work, please consider citing and star this repo:
@inproceedings{wang2021agreementmakerdeep,
title={AgreementMakerDeep results for OAEI 2021.},
author={Wang, Zhu and Cruz, Isabel F},
booktitle={OM@ ISWC},
pages={124--130},
year={2021}
}
and
@inproceedings{wang2022amd,
title={AMD results for OAEI 2022.},
author={Wang, Zhu},
booktitle={OM@ ISWC},
pages={145--152},
year={2022}
}