Conservative Policy Construction Using Variational Autoencoders for Logged Data With Missing Values

In high-stakes applications of data-driven decision-making such as healthcare, it is of paramount importance to learn a policy that maximizes the reward while avoiding potentially dangerous actions when there is uncertainty. There are two main challenges usually associated with this problem. First,...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transaction on neural networks and learning systems 2023-09, Vol.34 (9), p.6368-6378
Main Authors: Abroshan, Mahed, Yip, Kai Hou, Tekin, Cem, van der Schaar, Mihaela
Format: Article
Language:English
Subjects:
Datasets
Decision making
Estimation
IP networks
Missing values
Noise measurement
observational data
policy construction
Task analysis
Training data
Tuning
Uncertainty
variational autoencoder
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In high-stakes applications of data-driven decision-making such as healthcare, it is of paramount importance to learn a policy that maximizes the reward while avoiding potentially dangerous actions when there is uncertainty. There are two main challenges usually associated with this problem. First, learning through online exploration is not possible due to the critical nature of such applications. Therefore, we need to resort to observational datasets with no counterfactuals. Second, such datasets are usually imperfect, additionally cursed with missing values in the attributes of features. In this article, we consider the problem of constructing personalized policies using logged data when there are missing values in the attributes of features in both training and test data. The goal is to recommend an action (treatment) when \tilde { \boldsymbol {X}} , a degraded version of \boldsymbol {X} with missing values, is observed. We consider three strategies for dealing with missingness. In particular, we introduce the conservative strategy where the policy is designed to safely handle the uncertainty due to missingness. In order to implement this strategy, we need to estimate posterior distribution p(\boldsymbol {X}| \tilde { \boldsymbol {X}}) and use a variational autoencoder to achieve this. In particular, our method is based on partial variational autoencoders (PVAEs) that are designed to capture the underlying structure of features with missing values.
ISSN:2162-237X
2162-2388
DOI:10.1109/TNNLS.2021.3136385