Traversing Knowledge Graphs in Vector Space
This addresses the issue of cascading errors in answering compositional questions for users of knowledge graphs, representing a strong incremental improvement over existing embedding methods.
The paper tackled the problem of missing facts in knowledge graphs disrupting path queries by introducing a compositional training objective, which more than doubled accuracy for some models and reduced errors by up to 43% on a standard task, achieving new state-of-the-art results.
Path queries on a knowledge graph can be used to answer compositional questions such as "What languages are spoken by people living in Lisbon?". However, knowledge graphs often have missing facts (edges) which disrupts path queries. Recent models for knowledge base completion impute missing facts by embedding knowledge graphs in vector spaces. We show that these models can be recursively applied to answer path queries, but that they suffer from cascading errors. This motivates a new "compositional" training objective, which dramatically improves all models' ability to answer path queries, in some cases more than doubling accuracy. On a standard knowledge base completion task, we also demonstrate that compositional training acts as a novel form of structural regularization, reliably improving performance across all base models (reducing errors by up to 43%) and achieving new state-of-the-art results.