%0 Conference Paper %1 10.5555/3016100.3016273 %A Xie, Ruobing %A Liu, Zhiyuan %A Jia, Jia %A Luan, Huanbo %A Sun, Maosong %B Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence %D 2016 %I AAAI Press %K %P 2659–2665 %T Representation Learning of Knowledge Graphs with Entity Descriptions %X Representation learning (RL) of knowledge graphs aims to project both entities and relations into a continuous low-dimensional space. Most methods concentrate on learning representations with knowledge triples indicating relations between entities. In fact, in most knowledge graphs there are usually concise descriptions for entities, which cannot be well utilized by existing methods. In this paper, we propose a novel RL method for knowledge graphs taking advantages of entity descriptions. More specifically, we explore two encoders, including continuous bag-of-words and deep convolutional neural models to encode semantics of entity descriptions. We further learn knowledge representations with both triples and descriptions. We evaluate our method on two tasks, including knowledge graph completion and entity classification. Experimental results on real-world datasets show that, our method outperforms other baselines on the two tasks, especially under the zero-shot setting, which indicates that our method is capable of building representations for novel entities according to their descriptions. The source code of this paper can be obtained from https://github.com/xrb92/DKRL.

@inproceedings{10.5555/3016100.3016273, abstract = {Representation learning (RL) of knowledge graphs aims to project both entities and relations into a continuous low-dimensional space. Most methods concentrate on learning representations with knowledge triples indicating relations between entities. In fact, in most knowledge graphs there are usually concise descriptions for entities, which cannot be well utilized by existing methods. In this paper, we propose a novel RL method for knowledge graphs taking advantages of entity descriptions. More specifically, we explore two encoders, including continuous bag-of-words and deep convolutional neural models to encode semantics of entity descriptions. We further learn knowledge representations with both triples and descriptions. We evaluate our method on two tasks, including knowledge graph completion and entity classification. Experimental results on real-world datasets show that, our method outperforms other baselines on the two tasks, especially under the zero-shot setting, which indicates that our method is capable of building representations for novel entities according to their descriptions. The source code of this paper can be obtained from https://github.com/xrb92/DKRL.}, added-at = {2021-06-25T07:04:22.000+0200}, author = {Xie, Ruobing and Liu, Zhiyuan and Jia, Jia and Luan, Huanbo and Sun, Maosong}, biburl = {https://www.bibsonomy.org/bibtex/2186fd5821831218849684c55f8b57ca6/michan}, booktitle = {Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence}, interhash = {a1e56205d50a5efb684668ae21c74383}, intrahash = {186fd5821831218849684c55f8b57ca6}, keywords = {}, location = {Phoenix, Arizona}, numpages = {7}, pages = {2659–2665}, publisher = {AAAI Press}, series = {AAAI'16}, timestamp = {2021-06-25T07:04:22.000+0200}, title = {Representation Learning of Knowledge Graphs with Entity Descriptions}, year = 2016 }