VARIATIONAL AUTOENCODERS FOR HIGHLY MULTIVARIATE SPATIAL POINT PROCESSES INTENSITIES

Baichuan Yuan; Xiaowei Wang; Jianxin Ma; Chang Zhou; Andrea L. Bertozzi; Hongxia Yang

VARIATIONAL AUTOENCODERS FOR HIGHLY MULTIVARIATE SPATIAL POINT PROCESSES INTENSITIES

Baichuan Yuan, Xiaowei Wang, Jianxin Ma, Chang Zhou, Andrea L. Bertozzi, Hongxia Yang

Research output: Unpublished conference presentation (presented paper, abstract, poster) › Conference presentation (not published in journal/proceeding/book) › Academic research › peer-review

3 Citations (Scopus)

Abstract

Multivariate spatial point process models can describe heterotopic data over space. However, highly multivariate intensities are computationally challenging due to the curse of dimensionality. To bridge this gap, we introduce a declustering based hidden variable model that leads to an efficient inference procedure via a variational autoencoder (VAE). We also prove that this model is a generalization of the VAE-based model for collaborative filtering. This leads to an interesting application of spatial point process models to recommender systems. Experimental results show the method's utility on both synthetic data and real-world data sets.

Original language	English
Publication status	Published - Apr 2020
Externally published	Yes
Event	8th International Conference on Learning Representations, ICLR 2020 - Addis Ababa, Ethiopia Duration: 30 Apr 2020 → …

Conference

Conference	8th International Conference on Learning Representations, ICLR 2020
Country/Territory	Ethiopia
City	Addis Ababa
Period	30/04/20 → …

ASJC Scopus subject areas

Education
Linguistics and Language
Language and Linguistics
Computer Science Applications

Cite this

@conference{25255f3c89d34bfda7942c3b3ce9cf4a,

title = "VARIATIONAL AUTOENCODERS FOR HIGHLY MULTIVARIATE SPATIAL POINT PROCESSES INTENSITIES",

abstract = "Multivariate spatial point process models can describe heterotopic data over space. However, highly multivariate intensities are computationally challenging due to the curse of dimensionality. To bridge this gap, we introduce a declustering based hidden variable model that leads to an efficient inference procedure via a variational autoencoder (VAE). We also prove that this model is a generalization of the VAE-based model for collaborative filtering. This leads to an interesting application of spatial point process models to recommender systems. Experimental results show the method's utility on both synthetic data and real-world data sets.",

author = "Baichuan Yuan and Xiaowei Wang and Jianxin Ma and Chang Zhou and Bertozzi, \{Andrea L.\} and Hongxia Yang",

note = "Publisher Copyright: {\textcopyright} 2020 8th International Conference on Learning Representations, ICLR 2020. All rights reserved.; 8th International Conference on Learning Representations, ICLR 2020 ; Conference date: 30-04-2020",

year = "2020",

month = apr,

language = "English",

}

VARIATIONAL AUTOENCODERS FOR HIGHLY MULTIVARIATE SPATIAL POINT PROCESSES INTENSITIES. / Yuan, Baichuan; Wang, Xiaowei; Ma, Jianxin et al.
2020. Paper presented at 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia.

Research output: Unpublished conference presentation (presented paper, abstract, poster) › Conference presentation (not published in journal/proceeding/book) › Academic research › peer-review

TY - CONF

T1 - VARIATIONAL AUTOENCODERS FOR HIGHLY MULTIVARIATE SPATIAL POINT PROCESSES INTENSITIES

AU - Yuan, Baichuan

AU - Wang, Xiaowei

AU - Ma, Jianxin

AU - Zhou, Chang

AU - Bertozzi, Andrea L.

AU - Yang, Hongxia

PY - 2020/4

Y1 - 2020/4

N2 - Multivariate spatial point process models can describe heterotopic data over space. However, highly multivariate intensities are computationally challenging due to the curse of dimensionality. To bridge this gap, we introduce a declustering based hidden variable model that leads to an efficient inference procedure via a variational autoencoder (VAE). We also prove that this model is a generalization of the VAE-based model for collaborative filtering. This leads to an interesting application of spatial point process models to recommender systems. Experimental results show the method's utility on both synthetic data and real-world data sets.

AB - Multivariate spatial point process models can describe heterotopic data over space. However, highly multivariate intensities are computationally challenging due to the curse of dimensionality. To bridge this gap, we introduce a declustering based hidden variable model that leads to an efficient inference procedure via a variational autoencoder (VAE). We also prove that this model is a generalization of the VAE-based model for collaborative filtering. This leads to an interesting application of spatial point process models to recommender systems. Experimental results show the method's utility on both synthetic data and real-world data sets.

UR - https://www.scopus.com/pages/publications/85150626020

M3 - Conference presentation (not published in journal/proceeding/book)

AN - SCOPUS:85150626020

T2 - 8th International Conference on Learning Representations, ICLR 2020

Y2 - 30 April 2020

ER -

VARIATIONAL AUTOENCODERS FOR HIGHLY MULTIVARIATE SPATIAL POINT PROCESSES INTENSITIES

Abstract

Conference

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this