EDDI: Efficient Dynamic Discovery of High-Value Information with Partial VAE

Chao Ma, Sebastian Tschiatschek, Konstantina Palla, Jose Miguel Hernandez-Lobato, Sebastian Nowozin, Cheng Zhang
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:4234-4243, 2019.

Abstract

Many real-life decision making situations allow further relevant information to be acquired at a specific cost, for example, in assessing the health status of a patient we may decide to take additional measurements such as diagnostic tests or imaging scans before making a final assessment. Acquiring more relevant information enables better decision making, but may be costly. How can we trade off the desire to make good decisions by acquiring further information with the cost of performing that acquisition? To this end, we propose a principled framework, named EDDI (Efficient Dynamic Discovery of high-value Information), based on the theory of Bayesian experimental design. In EDDI, we propose a novel partial variational autoencoder (Partial VAE) to predict missing data entries problematically given any subset of the observed ones, and combine it with an acquisition function that maximizes expected information gain on a set of target variables. We show cost reduction at the same decision quality and improved decision quality at the same cost in multiple machine learning benchmarks and two real-world health-care applications.

Cite this Paper

BibTeX
@InProceedings{pmlr-v97-ma19c, title = {{EDDI}: Efficient Dynamic Discovery of High-Value Information with Partial {VAE}}, author = {Ma, Chao and Tschiatschek, Sebastian and Palla, Konstantina and Hernandez-Lobato, Jose Miguel and Nowozin, Sebastian and Zhang, Cheng}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {4234--4243}, 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/ma19c/ma19c.pdf}, url = {https://proceedings.mlr.press/v97/ma19c.html}, abstract = {Many real-life decision making situations allow further relevant information to be acquired at a specific cost, for example, in assessing the health status of a patient we may decide to take additional measurements such as diagnostic tests or imaging scans before making a final assessment. Acquiring more relevant information enables better decision making, but may be costly. How can we trade off the desire to make good decisions by acquiring further information with the cost of performing that acquisition? To this end, we propose a principled framework, named EDDI (Efficient Dynamic Discovery of high-value Information), based on the theory of Bayesian experimental design. In EDDI, we propose a novel partial variational autoencoder (Partial VAE) to predict missing data entries problematically given any subset of the observed ones, and combine it with an acquisition function that maximizes expected information gain on a set of target variables. We show cost reduction at the same decision quality and improved decision quality at the same cost in multiple machine learning benchmarks and two real-world health-care applications.} }
Endnote
%0 Conference Paper %T EDDI: Efficient Dynamic Discovery of High-Value Information with Partial VAE %A Chao Ma %A Sebastian Tschiatschek %A Konstantina Palla %A Jose Miguel Hernandez-Lobato %A Sebastian Nowozin %A Cheng Zhang %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-ma19c %I PMLR %P 4234--4243 %U https://proceedings.mlr.press/v97/ma19c.html %V 97 %X Many real-life decision making situations allow further relevant information to be acquired at a specific cost, for example, in assessing the health status of a patient we may decide to take additional measurements such as diagnostic tests or imaging scans before making a final assessment. Acquiring more relevant information enables better decision making, but may be costly. How can we trade off the desire to make good decisions by acquiring further information with the cost of performing that acquisition? To this end, we propose a principled framework, named EDDI (Efficient Dynamic Discovery of high-value Information), based on the theory of Bayesian experimental design. In EDDI, we propose a novel partial variational autoencoder (Partial VAE) to predict missing data entries problematically given any subset of the observed ones, and combine it with an acquisition function that maximizes expected information gain on a set of target variables. We show cost reduction at the same decision quality and improved decision quality at the same cost in multiple machine learning benchmarks and two real-world health-care applications.
APA
Ma, C., Tschiatschek, S., Palla, K., Hernandez-Lobato, J.M., Nowozin, S. & Zhang, C.. (2019). EDDI: Efficient Dynamic Discovery of High-Value Information with Partial VAE. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:4234-4243 Available from https://proceedings.mlr.press/v97/ma19c.html.

Related Material