Holographic Embeddings of Knowledge Graphs

Голографические вложения графов знаний
Lorenzo Rosasco, Tomaso Poggio, Maximilian Nickel
2016-03-02

circular correlationholographic embeddingsknowledge graphslink predictionrelational learning
Learning embeddings of entities and relations is an efficient and versatile method to perform machine learning on relational data such as knowledge graphs. In this work, we propose holographic embeddings (HolE) to learn compositional vector space representations of entire knowledge graphs. The proposed method is related to holographic models of associative memory in that it employs circular correlation to create compositional representations. By using correlation as the compositional operator, HolE can capture rich interactions but simultaneously remains efficient to compute, easy to train, and scalable to very large datasets. Experimentally, we show that holographic embeddings are able to outperform state-of-the-art methods for link prediction on knowledge graphs and relational learning benchmark datasets.
1
Experiments show that HolE outperforms state-of-the-art methods on knowledge-graph link prediction and relational-learning benchmarks.
2
HolE uses circular correlation as its composition operator, enabling rich interactions between entity and relation representations.
3
Holographic Embeddings (HolE) learn compositional vector representations for entire knowledge graphs.
4
The method remains computationally efficient, easy to train, and scalable to very large datasets.

Knowledge graphs (entities and relations represented for relational learning)

the effectiveness, expressive capacity, computational efficiency, trainability, scalability, and link-prediction performance of holographic embeddings for compositional representation of knowledge graphs

Publication Details
Publication Date
2016-03-02
Journal
Publisher
ISSN
Cited by
559
Access Type
Author Information
Authors
Lorenzo Rosasco
Tomaso Poggio
Maximilian Nickel
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%