Hyperbolic Disk Embeddings for Directed Acyclic Graphs

Ryota Suzuki, Ryusuke Takahama, Shun Onoda
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:6066-6075, 2019.

Abstract

Obtaining continuous representations of structural data such as directed acyclic graphs (DAGs) has gained attention in machine learning and artificial intelligence. However, embedding complex DAGs in which both ancestors and descendants of nodes are exponentially increasing is difficult. Tackling in this problem, we develop Disk Embeddings, which is a framework for embedding DAGs into quasi-metric spaces. Existing state-of-the-art methods, Order Embeddings and Hyperbolic Entailment Cones, are instances of Disk Embedding in Euclidean space and spheres respectively. Furthermore, we propose a novel method Hyperbolic Disk Embeddings to handle exponential growth of relations. The results of our experiments show that our Disk Embedding models outperform existing methods especially in complex DAGs other than trees.

Cite this Paper

BibTeX
@InProceedings{pmlr-v97-suzuki19a, title = {Hyperbolic Disk Embeddings for Directed Acyclic Graphs}, author = {Suzuki, Ryota and Takahama, Ryusuke and Onoda, Shun}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {6066--6075}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/suzuki19a/suzuki19a.pdf}, url = {https://proceedings.mlr.press/v97/suzuki19a.html}, abstract = {Obtaining continuous representations of structural data such as directed acyclic graphs (DAGs) has gained attention in machine learning and artificial intelligence. However, embedding complex DAGs in which both ancestors and descendants of nodes are exponentially increasing is difficult. Tackling in this problem, we develop Disk Embeddings, which is a framework for embedding DAGs into quasi-metric spaces. Existing state-of-the-art methods, Order Embeddings and Hyperbolic Entailment Cones, are instances of Disk Embedding in Euclidean space and spheres respectively. Furthermore, we propose a novel method Hyperbolic Disk Embeddings to handle exponential growth of relations. The results of our experiments show that our Disk Embedding models outperform existing methods especially in complex DAGs other than trees.} }
Endnote
%0 Conference Paper %T Hyperbolic Disk Embeddings for Directed Acyclic Graphs %A Ryota Suzuki %A Ryusuke Takahama %A Shun Onoda %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-suzuki19a %I PMLR %P 6066--6075 %U https://proceedings.mlr.press/v97/suzuki19a.html %V 97 %X Obtaining continuous representations of structural data such as directed acyclic graphs (DAGs) has gained attention in machine learning and artificial intelligence. However, embedding complex DAGs in which both ancestors and descendants of nodes are exponentially increasing is difficult. Tackling in this problem, we develop Disk Embeddings, which is a framework for embedding DAGs into quasi-metric spaces. Existing state-of-the-art methods, Order Embeddings and Hyperbolic Entailment Cones, are instances of Disk Embedding in Euclidean space and spheres respectively. Furthermore, we propose a novel method Hyperbolic Disk Embeddings to handle exponential growth of relations. The results of our experiments show that our Disk Embedding models outperform existing methods especially in complex DAGs other than trees.
APA
Suzuki, R., Takahama, R. & Onoda, S.. (2019). Hyperbolic Disk Embeddings for Directed Acyclic Graphs. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:6066-6075 Available from https://proceedings.mlr.press/v97/suzuki19a.html.

Related Material