Censor Dependent Variational Inference
This work addresses a problem in survival analysis, which is crucial for medical research and applications, providing an incremental yet significant improvement over existing variational inference methods.
The authors tackled the challenge of variational inference in latent variable models for survival analysis, and their proposed censor-dependent variational inference (CDVI) method achieved significant improvements in estimating individual survival distributions. The method addresses a critical weakness in existing methodology by accounting for the censoring mechanism.
This paper provides a comprehensive analysis of variational inference in latent variable models for survival analysis, emphasizing the distinctive challenges associated with applying variational methods to survival data. We identify a critical weakness in the existing methodology, demonstrating how a poorly designed variational distribution may hinder the objective of survival analysis tasks - modeling time-to-event distributions. We prove that the optimal variational distribution, which perfectly bounds the log-likelihood, may depend on the censoring mechanism. To address this issue, we propose censor-dependent variational inference (CDVI), tailored for latent variable models in survival analysis. More practically, we introduce CD-CVAE, a V-structure Variational Autoencoder (VAE) designed for the scalable implementation of CDVI. Further discussion extends some existing theories and training techniques to survival analysis. Extensive experiments validate our analysis and demonstrate significant improvements in the estimation of individual survival distributions.