%0 Conference Paper %1 demir2021shallow %A Demir, Caglar %A Moussallem, Diego %A Ngonga Ngomo, Axel-Cyrille %B Proceedings of the 15th IEEE International Conference on Semantic Computing (ICSC 2021) %D 2021 %K daikiri dice raki %T A shallow neural model for relation prediction %U https://papers.dice-research.org/2021/ICSC2021_Shallom/public.pdf %X Knowledge graph completion refers to predicting missing triples. Most approaches achieve this goal by predicting entities, given an entity and a relation. We predict missing triples via the relation prediction. To this end, we frame the relation prediction problem as a multi-label classification problem and propose a shallow neural model (SHALLOM) that accurately infers missing relations from entities. SHALLOM is analogous to C-BOW as both approaches predict a central token (p) given surrounding tokens ((s,o)). Our experiments indicate that SHALLOM outperforms state-of-the-art approaches on the FB15K-237 and WN18RR with margins of up to 3% and 8% (absolute), respectively, while requiring a maximum training time of 8 minutes on these datasets. We ensure the reproducibility of our results by providing an open-source implementation including training and evaluation scripts.

@inproceedings{demir2021shallow, abstract = {Knowledge graph completion refers to predicting missing triples. Most approaches achieve this goal by predicting entities, given an entity and a relation. We predict missing triples via the relation prediction. To this end, we frame the relation prediction problem as a multi-label classification problem and propose a shallow neural model (SHALLOM) that accurately infers missing relations from entities. SHALLOM is analogous to C-BOW as both approaches predict a central token (p) given surrounding tokens ((s,o)). Our experiments indicate that SHALLOM outperforms state-of-the-art approaches on the FB15K-237 and WN18RR with margins of up to 3% and 8% (absolute), respectively, while requiring a maximum training time of 8 minutes on these datasets. We ensure the reproducibility of our results by providing an open-source implementation including training and evaluation scripts.}, added-at = {2021-01-25T09:35:00.000+0100}, author = {Demir, Caglar and Moussallem, Diego and Ngonga Ngomo, Axel-Cyrille}, bdsk-url-1 = {https://papers.dice-research.org/2021/ICSC2021_Shallom/public.pdf}, biburl = {https://www.bibsonomy.org/bibtex/2db82dbbe5b9cf8ffeefa76c1a758b812/dice-research}, booktitle = {Proceedings of the 15th {IEEE} {International} {Conference} on {Semantic} {Computing} ({ICSC} 2021)}, interhash = {34f7d1340542f9ed7ae2a2e506f33d42}, intrahash = {db82dbbe5b9cf8ffeefa76c1a758b812}, keywords = {daikiri dice raki}, timestamp = {2023-04-25T16:34:07.000+0200}, title = {A shallow neural model for relation prediction}, url = {https://papers.dice-research.org/2021/ICSC2021_Shallom/public.pdf}, year = 2021 }