Dirichlet Simplex Nest and Geometric Inference

Mikhail Yurochkin, Aritra Guha, Yuekai Sun, Xuanlong Nguyen
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:7262-7271, 2019.

Abstract

We propose Dirichlet Simplex Nest, a class of probabilistic models suitable for a variety of data types, and develop fast and provably accurate inference algorithms by accounting for the model’s convex geometry and low dimensional simplicial structure. By exploiting the connection to Voronoi tessellation and properties of Dirichlet distribution, the proposed inference algorithm is shown to achieve consistency and strong error bound guarantees on a range of model settings and data distributions. The effectiveness of our model and the learning algorithm is demonstrated by simulations and by analyses of text and financial data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v97-yurochkin19b, title = {{D}irichlet Simplex Nest and Geometric Inference}, author = {Yurochkin, Mikhail and Guha, Aritra and Sun, Yuekai and Nguyen, Xuanlong}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {7262--7271}, 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/yurochkin19b/yurochkin19b.pdf}, url = {https://proceedings.mlr.press/v97/yurochkin19b.html}, abstract = {We propose Dirichlet Simplex Nest, a class of probabilistic models suitable for a variety of data types, and develop fast and provably accurate inference algorithms by accounting for the model’s convex geometry and low dimensional simplicial structure. By exploiting the connection to Voronoi tessellation and properties of Dirichlet distribution, the proposed inference algorithm is shown to achieve consistency and strong error bound guarantees on a range of model settings and data distributions. The effectiveness of our model and the learning algorithm is demonstrated by simulations and by analyses of text and financial data.} }
Endnote
%0 Conference Paper %T Dirichlet Simplex Nest and Geometric Inference %A Mikhail Yurochkin %A Aritra Guha %A Yuekai Sun %A Xuanlong Nguyen %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-yurochkin19b %I PMLR %P 7262--7271 %U https://proceedings.mlr.press/v97/yurochkin19b.html %V 97 %X We propose Dirichlet Simplex Nest, a class of probabilistic models suitable for a variety of data types, and develop fast and provably accurate inference algorithms by accounting for the model’s convex geometry and low dimensional simplicial structure. By exploiting the connection to Voronoi tessellation and properties of Dirichlet distribution, the proposed inference algorithm is shown to achieve consistency and strong error bound guarantees on a range of model settings and data distributions. The effectiveness of our model and the learning algorithm is demonstrated by simulations and by analyses of text and financial data.
APA
Yurochkin, M., Guha, A., Sun, Y. & Nguyen, X.. (2019). Dirichlet Simplex Nest and Geometric Inference. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:7262-7271 Available from https://proceedings.mlr.press/v97/yurochkin19b.html.

Related Material