Canonical Tensor Decomposition for Knowledge Base Completion

Timothee Lacroix, Nicolas Usunier, Guillaume Obozinski
Proceedings of the 35th International Conference on Machine Learning, PMLR 80:2863-2872, 2018.

Abstract

The problem of Knowledge Base Completion can be framed as a 3rd-order binary tensor completion problem. In this light, the Canonical Tensor Decomposition (CP) seems like a natural solution; however, current implementations of CP on standard Knowledge Base Completion benchmarks are lagging behind their competitors. In this work, we attempt to understand the limits of CP for knowledge base completion. First, we motivate and test a novel regularizer, based on tensor nuclear p-norms. Then, we present a reformulation of the problem that makes it invariant to arbitrary choices in the inclusion of predicates or their reciprocals in the dataset. These two methods combined allow us to beat the current state of the art on several datasets with a CP decomposition, and obtain even better results using the more advanced ComplEx model.

Cite this Paper

BibTeX
@InProceedings{pmlr-v80-lacroix18a, title = {Canonical Tensor Decomposition for Knowledge Base Completion}, author = {Lacroix, Timothee and Usunier, Nicolas and Obozinski, Guillaume}, booktitle = {Proceedings of the 35th International Conference on Machine Learning}, pages = {2863--2872}, year = {2018}, editor = {Dy, Jennifer and Krause, Andreas}, volume = {80}, series = {Proceedings of Machine Learning Research}, month = {10--15 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v80/lacroix18a/lacroix18a.pdf}, url = {https://proceedings.mlr.press/v80/lacroix18a.html}, abstract = {The problem of Knowledge Base Completion can be framed as a 3rd-order binary tensor completion problem. In this light, the Canonical Tensor Decomposition (CP) seems like a natural solution; however, current implementations of CP on standard Knowledge Base Completion benchmarks are lagging behind their competitors. In this work, we attempt to understand the limits of CP for knowledge base completion. First, we motivate and test a novel regularizer, based on tensor nuclear p-norms. Then, we present a reformulation of the problem that makes it invariant to arbitrary choices in the inclusion of predicates or their reciprocals in the dataset. These two methods combined allow us to beat the current state of the art on several datasets with a CP decomposition, and obtain even better results using the more advanced ComplEx model.} }
Endnote
%0 Conference Paper %T Canonical Tensor Decomposition for Knowledge Base Completion %A Timothee Lacroix %A Nicolas Usunier %A Guillaume Obozinski %B Proceedings of the 35th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2018 %E Jennifer Dy %E Andreas Krause %F pmlr-v80-lacroix18a %I PMLR %P 2863--2872 %U https://proceedings.mlr.press/v80/lacroix18a.html %V 80 %X The problem of Knowledge Base Completion can be framed as a 3rd-order binary tensor completion problem. In this light, the Canonical Tensor Decomposition (CP) seems like a natural solution; however, current implementations of CP on standard Knowledge Base Completion benchmarks are lagging behind their competitors. In this work, we attempt to understand the limits of CP for knowledge base completion. First, we motivate and test a novel regularizer, based on tensor nuclear p-norms. Then, we present a reformulation of the problem that makes it invariant to arbitrary choices in the inclusion of predicates or their reciprocals in the dataset. These two methods combined allow us to beat the current state of the art on several datasets with a CP decomposition, and obtain even better results using the more advanced ComplEx model.
APA
Lacroix, T., Usunier, N. & Obozinski, G.. (2018). Canonical Tensor Decomposition for Knowledge Base Completion. Proceedings of the 35th International Conference on Machine Learning, in Proceedings of Machine Learning Research 80:2863-2872 Available from https://proceedings.mlr.press/v80/lacroix18a.html.

Related Material