%0 Conference Paper %1 10.1007/978-3-031-43458-7_50 %A Chudasama, Yashrajsinh %B The Semantic Web: ESWC 2023 Satellite Events %C Cham %D 2023 %E Pesquita, Catia %E Skaf-Molli, Hala %E Efthymiou, Vasilis %E Kirrane, Sabrina %E Ngonga, Axel %E Collarana, Diego %E Cerqueira, Renato %E Alam, Mehwish %E Trojahn, Cassia %E Hertling, Sven %I Springer Nature Switzerland %K myown %P 321--330 %R https://doi.org/10.1007/978-3-031-43458-7_50 %T Exploiting Semantics for Explaining Link Prediction Over Knowledge Graphs %X The use of Symbolic and sub-symbolic AI techniques on Knowledge Graphs (KGs) has shown significant progress in several applications. However, many of these methods remain opaque, and the decision-making process behind them can be perplexing. This can result in a lack of trust and reliability in the overall framework. While various explainable frameworks have been proposed to address these issues, do not always provide a complete understanding and may raise privacy concerns as sensitive data may be revealed during the explanation process. In contrast, our proposed approach leverages the semantics of KGs and causal relationships to enhance explainability while still maintaining a high level of trust and reliability. By focusing on XAI for link prediction models and considering entailment regimes (e.g., rdfs:subPropertyOf), the approach can provide more comprehensive and accurate explanations. Moreover, the use of symbolic reasoning allows for more transparent and interpretable explanations. The preliminary results show that our approach is capable of exploiting the semantics of an entity in KG and enhancing the explanations. Henceforth, more work needs to be conducted, to fully comprehend all impacting factors and to identify the most relevant explanations of the machine learning models over KGs. %@ 978-3-031-43458-7

@inproceedings{10.1007/978-3-031-43458-7_50, abstract = {The use of Symbolic and sub-symbolic AI techniques on Knowledge Graphs (KGs) has shown significant progress in several applications. However, many of these methods remain opaque, and the decision-making process behind them can be perplexing. This can result in a lack of trust and reliability in the overall framework. While various explainable frameworks have been proposed to address these issues, do not always provide a complete understanding and may raise privacy concerns as sensitive data may be revealed during the explanation process. In contrast, our proposed approach leverages the semantics of KGs and causal relationships to enhance explainability while still maintaining a high level of trust and reliability. By focusing on XAI for link prediction models and considering entailment regimes (e.g., rdfs:subPropertyOf), the approach can provide more comprehensive and accurate explanations. Moreover, the use of symbolic reasoning allows for more transparent and interpretable explanations. The preliminary results show that our approach is capable of exploiting the semantics of an entity in KG and enhancing the explanations. Henceforth, more work needs to be conducted, to fully comprehend all impacting factors and to identify the most relevant explanations of the machine learning models over KGs.}, added-at = {2024-02-15T15:11:01.000+0100}, address = {Cham}, author = {Chudasama, Yashrajsinh}, biburl = {https://www.bibsonomy.org/bibtex/2469aaa1d16a54b70de1d0ac87c3e9e9e/gabydler}, booktitle = {The Semantic Web: ESWC 2023 Satellite Events}, doi = {https://doi.org/10.1007/978-3-031-43458-7_50}, editor = {Pesquita, Catia and Skaf-Molli, Hala and Efthymiou, Vasilis and Kirrane, Sabrina and Ngonga, Axel and Collarana, Diego and Cerqueira, Renato and Alam, Mehwish and Trojahn, Cassia and Hertling, Sven}, interhash = {2bef2bafcd5ca965580b27831a9aebab}, intrahash = {469aaa1d16a54b70de1d0ac87c3e9e9e}, isbn = {978-3-031-43458-7}, keywords = {myown}, pages = {321--330}, publisher = {Springer Nature Switzerland}, timestamp = {2024-02-15T15:11:01.000+0100}, title = {Exploiting Semantics for Explaining Link Prediction Over Knowledge Graphs}, year = 2023 }