Filters








188,008 Hits in 4.6 sec

Interpretable Graph Convolutional Neural Networks for Inference on Noisy Knowledge Graphs [article]

Daniel Neil, Joss Briody, Alix Lacoste, Aaron Sim, Paidi Creed, Amir Saffari
2018 arXiv   pre-print
In this work, we provide a new formulation for Graph Convolutional Neural Networks (GCNNs) for link prediction on graph data that addresses common challenges for biomedical knowledge graphs (KGs).  ...  The results are demonstrated on a synthetic dataset, the common benchmark dataset FB15k-237, and a large biomedical knowledge graph derived from a combination of noisy and clean data sources.  ...  Correctness and Interpretation in a Biological Knowledge Graph As an example real-world application, we examine a subset of a proprietary knowledge graph with 708k edges, compiled from unstructured data  ... 
arXiv:1812.00279v1 fatcat:wsjcevg3rvdrtldq5wbmmqwpdu

KGRefiner: Knowledge Graph Refinement for Improving Accuracy of Translational Link Prediction Methods [article]

Mohammad Javad Saeedizade, Najmeh Torabian, Behrouz Minaei-Bidgoli
2021 arXiv   pre-print
The Link Prediction is the task of predicting missing relations between entities of the knowledge graph.  ...  In this paper, we propose a novel method of refining the knowledge graph so that link prediction operation can be performed more accurately using relatively fast translational models.  ...  One application of these embedding is to predict missing links in the knowledge graph. Translational link prediction models use the sum of the head and relation vectors to predict the tail.  ... 
arXiv:2106.14233v2 fatcat:irgfoeht3ja3hpp55aate2fkly

Neural Bellman-Ford Networks: A General Graph Neural Network Framework for Link Prediction [article]

Zhaocheng Zhu, Zuobai Zhang, Louis-Pascal Xhonneux, Jian Tang
2022 arXiv   pre-print
Link prediction is a very fundamental task on graphs.  ...  Inspired by traditional path-based methods, in this paper we propose a general and flexible representation learning framework based on paths for link prediction.  ...  Acknowledgements We would like to thank Komal Teru for discussion on inductive relation prediction, Guyue Huang for discussion on fused message passing implementation, and Yao Lu for assistance on large-scale  ... 
arXiv:2106.06935v4 fatcat:mnxi6uevp5g77pe7nkdc5aquqa

Graph Pattern Entity Ranking Model for Knowledge Graph Completion [article]

Takuma Ebisu, Ryutaro Ichise
2019 arXiv   pre-print
By doing so, we can find graph patterns which are useful for predicting facts. Then, we perform link prediction tasks on standard datasets to evaluate our GRank method.  ...  However, knowledge graph embedding models are so-called black boxes, and the user does not know how the information in a knowledge graph is processed and the models can be difficult to interpret.  ...  These models are not only easy to interpret compared to knowledge graph embedding models but also outperform state-of-the-art models for link prediction.  ... 
arXiv:1904.02856v1 fatcat:3rvnhhrb4zc47gtbukrvatr3sy

Investigating Robustness and Interpretability of Link Prediction via Adversarial Modifications [article]

Pouya Pezeshkpour, Yifan Tian, Sameer Singh
2019 arXiv   pre-print
In this paper, we propose adversarial modifications for link prediction models: identifying the fact to add into or remove from the knowledge graph that changes the prediction for a target fact after the  ...  We use these techniques to evaluate the robustness of link prediction models (by measuring sensitivity to additional facts), study interpretability through the facts most responsible for predictions (by  ...  Interpretability of Models To be able to understand and interpret why a link is predicted using the opaque, dense embeddings, we need to find out which part of the graph was most influential on the prediction  ... 
arXiv:1905.00563v1 fatcat:bv6266jjsvg4hisz5pladzopzm

Investigating Robustness and Interpretability of Link Prediction via Adversarial Modifications

Pouya Pezeshkpour, Yifan Tian, Sameer Singh
2019 Proceedings of the 2019 Conference of the North  
In this paper, we propose adversarial modifications for link prediction models: identifying the fact to add into or remove from the knowledge graph that changes the prediction for a target fact after the  ...  We use these techniques to evaluate the robustness of link prediction models (by measuring sensitivity to additional facts), study interpretability through the facts most responsible for predictions (by  ...  Interpretability of Models To be able to understand and interpret why a link is predicted using the opaque, dense embeddings, we need to find out which part of the graph was most influential on the prediction  ... 
doi:10.18653/v1/n19-1337 dblp:conf/naacl/PezeshkpourT019 fatcat:nu5xwjvyjvahlaia4ycgsrxk6a

Towards Adversarially Robust Knowledge Graph Embeddings

Peru Bhardwaj
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Knowledge graph embedding models enable representation learning on multi-relational graphs and are used in security sensitive domains. But, their security analysis has received little attention.  ...  I will research security of these models by designing adversarial attacks against them, improving their adversarial robustness and evaluating the effect of proposed improvement on their interpretability  ...  Embeddings learned from graphs can be used to classify nodes, to predict links or as background knowledge to other machine learning (ML) tasks.  ... 
doi:10.1609/aaai.v34i10.7128 fatcat:p5jqblu7kbbjtdks5qjsuiloki

Mandolin: A Knowledge Discovery Framework for the Web of Data [article]

Tommaso Soru, Diego Esteves, Edgard Marx, Axel-Cyrille Ngonga Ngomo
2017 arXiv   pre-print
We show that our best configuration scales well and achieves at least comparable results with respect to other statistical-relational-learning algorithms on link prediction.  ...  Our framework imports knowledge from referenced graphs, creates similarity relationships among similar literals, and relies on state-of-the-art techniques for rule mining, grounding, and inference computation  ...  Moreover, it can predict equivalence links across datasets and scale on large graphs.  ... 
arXiv:1711.01283v1 fatcat:qrzlefmn5jgeth554vt4olwiku

Data Augmentation for Personal Knowledge Base Population [article]

Lingraj S Vannur, Balaji Ganesan, Lokesh Nagalapatti, Hima Patel, MN Thippeswamy
2020 arXiv   pre-print
In this work, we present a system that uses rule based annotators and a graph neural network for missing link prediction, to populate a more complete, fair and diverse knowledge base from the TACRED dataset  ...  Cold start knowledge base population (KBP) is the problem of populating a knowledge base from unstructured documents.  ...  Link Prediction As shown in 2, Position Aware Graph Neural Network performs better than Graph Convolutional Networks on the TACRED Dataset.  ... 
arXiv:2002.10943v2 fatcat:nub7jctfkvbz5atv3o2h6xj2he

RatE: Relation-Adaptive Translating Embedding for Knowledge Graph Completion [article]

Hao Huang, Guodong Long, Tao Shen, Jing Jiang, Chengqi Zhang
2020 arXiv   pre-print
Many graph embedding approaches have been proposed for knowledge graph completion via link prediction.  ...  Experiments verify RatE achieves state-of-the-art performance on four link prediction benchmarks.  ...  Evaluation on Link Prediction Link prediction results on the four datasets are shown in Table 3 and Table 4 .  ... 
arXiv:2010.04863v1 fatcat:b3a5a4vxhjbgxm3nmcgdqwlzee

ELPKG: A High-Accuracy Link Prediction Approach for Knowledge Graph Completion

Jiangtao Ma, Yaqiong Qiao, Guangwu Hu, Yanjun Wang, Chaoqin Zhang, Yongzhong Huang, Arun Kumar Sangaiah, Huaiguang Wu, Hongpo Zhang, Kai Ren
2019 Symmetry  
To improve the accuracy of link predication, we propose a novel approach named Entity Link Prediction for Knowledge Graph (ELPKG), which can achieve a high accuracy on large-scale knowledge graphs while  ...  Link prediction in knowledge graph is the task of utilizing the existing relations to infer new relations so as to build a more complete knowledge graph.  ...  Acknowledgments: The authors would like to thank the editors and the reviewers for their comments on an earlier draft of this article.  ... 
doi:10.3390/sym11091096 fatcat:xiju3svx65d7ppjhp2m4b7au3y

Trust from the past: Bayesian Personalized Ranking based Link Prediction in Knowledge Graphs [article]

Baichuan Zhang, Sutanay Choudhury, Mohammad Al Hasan, Xia Ning, Khushbu Agarwal, Sumit Purohit, Paola Pesntez Cabrera
2016 arXiv   pre-print
Link prediction, or predicting the likelihood of a link in a knowledge graph based on its existing state is a key research task.  ...  It differs from a traditional link prediction task in that the links in a knowledge graph are categorized into different predicates and the link prediction performance of different predicates in a knowledge  ...  Related Work There is a large body of work on link prediction in knowledge graph.  ... 
arXiv:1601.03778v2 fatcat:unzkb7ifbzdgtcni23k2qbxfzy

Explainable Knowledge Graph Embedding: Inference Reconciliation for Knowledge Inferences Supporting Robot Actions [article]

Angel Daruna, Devleena Das, Sonia Chernova
2022 arXiv   pre-print
Our interpretable model, uses a decision tree classifier to locally approximate the predictions of the black-box model, and provides natural language explanations interpretable by non-experts.  ...  We use a pedagogical approach to explain the inferences of a learned, black-box knowledge graph representation, a knowledge graph embedding.  ...  We train the graph feature model to locally approximate the predictions of the learned knowledge graph representation and provide a grounded, natural language explanation for each prediction.  ... 
arXiv:2205.01836v1 fatcat:d65fod3h2ba4rn7nmog5gn26om

Combination of Unified Embedding Model and Observed Features for Knowledge Graph Completion [article]

Takuma Ebisu, Ryutaro Ichise
2019 arXiv   pre-print
Knowledge graphs are useful for many artificial intelligence tasks but often have missing data. Hence, a method for completing knowledge graphs is required.  ...  Then, we show that these models utilize paths for link prediction and propose a method for evaluating rules based on this idea.  ...  Several approaches have been proposed for link prediction, such as knowledge graph embedding models, rule evaluation models, and the Path Ranking Algorithm (PRA).  ... 
arXiv:1909.03821v2 fatcat:y4nh35oej5hwheir4rg5nves7m

Towards a Knowledge Graph based Speech Interface [article]

Ashwini Jaya Kumar, Sören Auer, Christoph Schmidt, Joachim köhler
2017 arXiv   pre-print
DBpedia Spotlight, a tool to interlink text documents with the linked open data is used to link the speech recognition output to the DBpedia knowledge graph.  ...  We show that for a corpus with lower WER, the annotation and linking of entities to the DBpedia knowledge graph is considerable.  ...  Linking Speech to Knowledge Graphs A speech interface to applications with speech as input involves recognising and interpreting the spoken utterance.  ... 
arXiv:1705.09222v1 fatcat:ilykxjfhy5a3popspb63k3lw5i
« Previous Showing results 1 — 15 out of 188,008 results