Scalable Structure Discovery in Regression using Gaussian Processes

Hyunjik Kim, Yee Whye Teh
Proceedings of the Workshop on Automatic Machine Learning, PMLR 64:31-40, 2016.

Abstract

Automatic Bayesian Covariance Discovery (ABCD) in Lloyd et. al (2014) provides a framework for automating statistical modelling as well as exploratory data analysis for regression problems. However ABCD does not scale due to its O(N3) running time. This is undesirable not only because the average size of data sets is growing fast, but also because there is potentially more information in bigger data, implying a greater need for more expressive models that can discover sophisticated structure. We propose a scalable version of ABCD, to encompass big data within the boundaries of automated statistical modelling.

Cite this Paper


BibTeX
@InProceedings{pmlr-v64-kim_scalable_2016, title = {Scalable Structure Discovery in Regression using Gaussian Processes}, author = {Kim, Hyunjik and Teh, Yee Whye}, booktitle = {Proceedings of the Workshop on Automatic Machine Learning}, pages = {31--40}, year = {2016}, editor = {Hutter, Frank and Kotthoff, Lars and Vanschoren, Joaquin}, volume = {64}, series = {Proceedings of Machine Learning Research}, address = {New York, New York, USA}, month = {24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v64/kim_scalable_2016.pdf}, url = {https://proceedings.mlr.press/v64/kim_scalable_2016.html}, abstract = {Automatic Bayesian Covariance Discovery (ABCD) in Lloyd et. al (2014) provides a framework for automating statistical modelling as well as exploratory data analysis for regression problems. However ABCD does not scale due to its $O(N^3)$ running time. This is undesirable not only because the average size of data sets is growing fast, but also because there is potentially more information in bigger data, implying a greater need for more expressive models that can discover sophisticated structure. We propose a scalable version of ABCD, to encompass big data within the boundaries of automated statistical modelling.} }
Endnote
%0 Conference Paper %T Scalable Structure Discovery in Regression using Gaussian Processes %A Hyunjik Kim %A Yee Whye Teh %B Proceedings of the Workshop on Automatic Machine Learning %C Proceedings of Machine Learning Research %D 2016 %E Frank Hutter %E Lars Kotthoff %E Joaquin Vanschoren %F pmlr-v64-kim_scalable_2016 %I PMLR %P 31--40 %U https://proceedings.mlr.press/v64/kim_scalable_2016.html %V 64 %X Automatic Bayesian Covariance Discovery (ABCD) in Lloyd et. al (2014) provides a framework for automating statistical modelling as well as exploratory data analysis for regression problems. However ABCD does not scale due to its $O(N^3)$ running time. This is undesirable not only because the average size of data sets is growing fast, but also because there is potentially more information in bigger data, implying a greater need for more expressive models that can discover sophisticated structure. We propose a scalable version of ABCD, to encompass big data within the boundaries of automated statistical modelling.
RIS
TY - CPAPER TI - Scalable Structure Discovery in Regression using Gaussian Processes AU - Hyunjik Kim AU - Yee Whye Teh BT - Proceedings of the Workshop on Automatic Machine Learning DA - 2016/12/04 ED - Frank Hutter ED - Lars Kotthoff ED - Joaquin Vanschoren ID - pmlr-v64-kim_scalable_2016 PB - PMLR DP - Proceedings of Machine Learning Research VL - 64 SP - 31 EP - 40 L1 - http://proceedings.mlr.press/v64/kim_scalable_2016.pdf UR - https://proceedings.mlr.press/v64/kim_scalable_2016.html AB - Automatic Bayesian Covariance Discovery (ABCD) in Lloyd et. al (2014) provides a framework for automating statistical modelling as well as exploratory data analysis for regression problems. However ABCD does not scale due to its $O(N^3)$ running time. This is undesirable not only because the average size of data sets is growing fast, but also because there is potentially more information in bigger data, implying a greater need for more expressive models that can discover sophisticated structure. We propose a scalable version of ABCD, to encompass big data within the boundaries of automated statistical modelling. ER -
APA
Kim, H. & Teh, Y.W.. (2016). Scalable Structure Discovery in Regression using Gaussian Processes. Proceedings of the Workshop on Automatic Machine Learning, in Proceedings of Machine Learning Research 64:31-40 Available from https://proceedings.mlr.press/v64/kim_scalable_2016.html.

Related Material