Adversarial Learning for Debiasing Knowledge Graph Embeddings

Knowledge Graphs (KG) are gaining increasing attention in both academia and industry. Despite their diverse benefits, recent research have identified social and cultural biases embedded in the representations learned from KGs. Such biases can have detrimental consequences on different population and minority groups as applications of KG begin to intersect and interact with social spheres. This paper aims at identifying and mitigating such biases in Knowledge Graph (KG) embeddings. As a first step, we explore popularity bias -- the relationship between node popularity and link prediction accuracy. In case of node2vec graph embeddings, we find that prediction accuracy of the embedding is negatively correlated with the degree of the node. However, in case of knowledge-graph embeddings (KGE), we observe an opposite trend. As a second step, we explore gender bias in KGE, and a careful examination of popular KGE algorithms suggest that sensitive attribute like the gender of a person can be predicted from the embedding. This implies that such biases in popular KGs is captured by the structural properties of the embedding. As a preliminary solution to debiasing KGs, we introduce a novel framework to filter out the sensitive attribute information from the KG embeddings, which we call FAN (Filtering Adversarial Network). We also suggest the applicability of FAN for debiasing other network embeddings which could be explored in future work.

[1]  Guillaume Bouchard,et al.  Complex Embeddings for Simple Link Prediction , 2016, ICML.

[2]  Jun Zhao,et al.  Knowledge Graph Embedding via Dynamic Mapping Matrix , 2015, ACL.

[3]  Jeffrey Dean,et al.  Efficient Estimation of Word Representations in Vector Space , 2013, ICLR.

[4]  Arvind Narayanan,et al.  Semantics derived automatically from language corpora contain human-like biases , 2016, Science.

[5]  Jens Lehmann,et al.  DBpedia: A Nucleus for a Web of Open Data , 2007, ISWC/ASWC.

[6]  Steven Skiena,et al.  DeepWalk: online learning of social representations , 2014, KDD.

[7]  Kalina Bontcheva,et al.  Named Entity Disambiguation using Linked Data , 2012 .

[8]  Wei Zhang,et al.  Interaction Embeddings for Prediction and Explanation in Knowledge Graphs , 2019, WSDM.

[9]  Andrew Chou,et al.  Semantic Parsing on Freebase from Question-Answer Pairs , 2013, EMNLP.

[10]  Hans-Peter Kriegel,et al.  A Three-Way Model for Collective Learning on Multi-Relational Data , 2011, ICML.

[11]  Yongjie Yang,et al.  Fairness in Algorithmic Decision-Making: Applications in Multi-Winner Voting, Machine Learning, and Recommender Systems , 2019, Algorithms.

[12]  Lorenzo Rosasco,et al.  Holographic Embeddings of Knowledge Graphs , 2015, AAAI.

[13]  Edward Y. Chang,et al.  Entity Disambiguation with Freebase , 2012, 2012 IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology.

[14]  Jason Weston,et al.  Open Question Answering with Weakly Supervised Embedding Models , 2014, ECML/PKDD.

[15]  Evgeny Kharlamov,et al.  Learning Rules from Incomplete KGs using Embeddings , 2018, SEMWEB.

[16]  Jason Weston,et al.  Question Answering with Subgraph Embeddings , 2014, EMNLP.

[17]  Gökhan Tür,et al.  Leveraging knowledge graphs for web-scale unsupervised semantic parsing , 2013, INTERSPEECH.

[18]  Zhen Wang,et al.  Knowledge Graph Embedding by Translating on Hyperplanes , 2014, AAAI.

[19]  Krzysztof Janowicz,et al.  Debiasing Knowledge Graphs: Why Female Presidents are not like Female Popes , 2018, SEMWEB.

[20]  Zhiyuan Liu,et al.  OpenKE: An Open Toolkit for Knowledge Embedding , 2018, EMNLP.

[21]  Evgeny Kharlamov,et al.  Rule Learning from Knowledge Graphs Guided by Embedding Models , 2018, SEMWEB.

[22]  Jure Leskovec,et al.  node2vec: Scalable Feature Learning for Networks , 2016, KDD.

[23]  Jason Weston,et al.  Translating Embeddings for Modeling Multi-relational Data , 2013, NIPS.

[24]  Abraham Bernstein,et al.  Updatable, Accurate, Diverse, and Scalable Recommendations for Interactive Applications , 2016, ACM Trans. Interact. Intell. Syst..

[25]  Gianluca Demartini,et al.  Implicit Bias in Crowdsourced Knowledge Graphs , 2019, WWW.

[26]  Jianfeng Gao,et al.  Embedding Entities and Relations for Learning and Inference in Knowledge Bases , 2014, ICLR.

[27]  William L. Hamilton,et al.  Compositional Fairness Constraints for Graph Embeddings , 2019, ICML.