Interpreting embedding models of knowledge bases.

Knowledge bases are employed in a variety of applications, from natural language processing to semantic web search; alas, in practice, their usefulness is hurt by their incompleteness. To address this issue, several techniques aim at performing knowledge base completion, of which embedding models are efficient, attain state-of-the-art accuracy, and eliminate the need for feature engineering. However, embedding models predictions are notoriously hard to interpret. In this work, we propose model-agnostic methods that allow one to interpret embedding models by extracting weighted Horn rules from them. More specifically, we show how the so-called “pedagogical techniques”, from the literature on neural networks, can be adapted to take into account the large-scale relational aspects of knowledge bases, and show experimentally their strengths and weaknesses.

• Publication year

  2018

• Venue

  Unknown venue

• Publication date

  2018-11-26

• Fields of study

  Computer Science

• Identifiers
  DOI 10.11606/D.3.2019.TDE-04022019-094854
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• OpenKE: An Open Toolkit for Knowledge Embedding

  2018cited by this paper
• An Interpretable Knowledge Transfer Model for Knowledge Base Completion

  2017cited by this paper
• An overview of embedding models of entities and relationships for knowledge base completion

  2017cited by this paper
• Knowledge Graph Embedding: A Survey of Approaches and Applications

  2017influential reference
• Mode-directed Neural-Symbolic Modelling

  2017cited by this paper
• What Does Explainable AI Really Mean? A New Conceptualization of Perspectives

  2017cited by this paper
• “Why Should I Trust You?”: Explaining the Predictions of Any Classifier

  2016cited by this paper
• Neighborhood Mixture Model for Knowledge Base Completion

  2016cited by this paper
• KSR: A Semantic Representation of Knowledge Graph within a Novel Unsupervised Paradigm

  2016cited by this paper
• The mythos of model interpretability

  2016cited by this paper
• Explainable and Efficient Link Prediction in Real-World Network Data

  2016cited by this paper
• Complex Embeddings for Simple Link Prediction

  2016cited by this paper
• DeepRED - Rule Extraction from Deep Neural Networks

  2016cited by this paper
• Semantically Smooth Knowledge Graph Embedding

  2015cited by this paper
• Neural-Symbolic Learning and Reasoning: Contributions and Challenges

  2015cited by this paper
• Type-Constrained Representation Learning in Knowledge Graphs

  2015cited by this paper
• Combining Vector Space Embeddings with Symbolic Logical Inference over Open-Domain Text

  2015cited by this paper
• Evaluating Link Prediction on Large Graphs

  2015influential reference
• Extracting Interpretable Models from Matrix Factorization Models

  2015cited by this paper
• A Review of Relational Machine Learning for Knowledge Graphs

  2015influential reference
• Never-Ending Learning

  2015cited by this paper
• Relational Knowledge Extraction from Neural Networks

  2015cited by this paper
• Modeling Relation Paths for Representation Learning of Knowledge Bases

  2015cited by this paper
• Efficient and Expressive Knowledge Base Completion Using Subgraph Feature Extraction

  2015cited by this paper
• Embedding Methods for Natural Language Processing

  2014cited by this paper
• Comprehensible classification models: a position paper

  2014cited by this paper
• Knowledge-based graph document modeling

  2014cited by this paper
• Knowledge Graph Embedding by Translating on Hyperplanes

  2014influential reference
• Who to follow and why: link prediction with explanations

  2014cited by this paper
• Efficient Estimation of Word Representations in Vector Space

  2013cited by this paper
• Translating Embeddings for Modeling Multi-relational Data

  2013cited by this paper
• A semantic matching energy function for learning with multi-relational data

  2013cited by this paper
• Reverse Engineering the Neural Networks for Rule Extraction in Classification Problems

  2012cited by this paper
• Learning Effective and Interpretable Semantic Models using Non-Negative Sparse Embedding

  2012cited by this paper
• Random Walk Inference and Learning in A Large Scale Knowledge Base

  2011cited by this paper
• A Three-Way Model for Collective Learning on Multi-Relational Data

  2011cited by this paper
• Relational retrieval using a combination of path-constrained random walks

  2010cited by this paper
• Life sciences on the Semantic Web: the Neurocommons and beyond

  2009cited by this paper
• Logical and relational learning

  2008cited by this paper
• Freebase: a collaboratively created graph database for structuring human knowledge

  2008cited by this paper
• Large-Scale Named Entity Disambiguation Based on Wikipedia Data

  2007cited by this paper
• Introduction to Statistical Relational Learning

  2007cited by this paper
• Orthogonal search-based rule extraction (OSRE) for trained neural networks: a practical and efficient approach

  2006cited by this paper
• WordNet: A Lexical Database for English

  1995cited by this paper
• Survey and critique of techniques for extracting rules from trained artificial neural networks

  1995cited by this paper
• in Advances in Neural Information Processing

  1989cited by this paper

• No citing papers are available for this paper.