Timezone: »

 
Poster
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

Wed Jun 12 06:30 PM -- 09:00 PM (PDT) @ Pacific Ballroom #6
in Posters Wed »

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 \emph{EDDI} (Efficient Dynamic Discovery of high-value Information), based on the theory of Bayesian experimental design. In EDDI, we propose a novel \emph{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.

Keywords: Bayesian Deep Learning · Deep Generative Models · Generative Models · Healthcare

Author Information

Chao Ma (University of Cambridge)
Sebastian Tschiatschek (Microsoft Research)
Konstantina Palla (Microsoft Research Cambridge)
Jose Miguel Hernandez-Lobato (University of Cambridge)
Sebastian Nowozin (MSR Cambridge)
Cheng Zhang (Microsoft Research, Cambridge)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors