DAGSurv: Directed Ayclic Graph Based Survival Analysis Using Deep Neural Networks

Ansh Kumar Sharma, Rahul Kukreja, Ranjitha Prasad, Shilpa Rao
Proceedings of The 13th Asian Conference on Machine Learning, PMLR 157:1065-1080, 2021.

Abstract

Causal structures for observational survival data provide crucial information regarding the relationships between covariates and time-to-event. We derive motivation from the information theoretic source coding argument, and show that incorporating the knowledge of the directed acyclic graph (DAG) can be beneficial if suitable source encoders are employed. As a possible source encoder in this context, we derive a variational inference based conditional variational autoencoder for causal structured survival prediction, which we refer to as \texttt{DAGSurv}. We illustrate the performance of \texttt{DAGSurv} on low and high-dimensional synthetic datasets, and real-world datasets such as METABRIC and GBSG. We demonstrate that the proposed method outperforms other survival analysis baselines such as \texttt{Cox} Proportional Hazards, \texttt{DeepSurv} and \texttt{Deephit}, which are oblivious to the underlying causal relationship between data entities.

Cite this Paper


BibTeX
@InProceedings{pmlr-v157-sharma21a, title = {DAGSurv: Directed Ayclic Graph Based Survival Analysis Using Deep Neural Networks}, author = {Sharma, Ansh Kumar and Kukreja, Rahul and Prasad, Ranjitha and Rao, Shilpa}, booktitle = {Proceedings of The 13th Asian Conference on Machine Learning}, pages = {1065--1080}, year = {2021}, editor = {Balasubramanian, Vineeth N. and Tsang, Ivor}, volume = {157}, series = {Proceedings of Machine Learning Research}, month = {17--19 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v157/sharma21a/sharma21a.pdf}, url = {https://proceedings.mlr.press/v157/sharma21a.html}, abstract = {Causal structures for observational survival data provide crucial information regarding the relationships between covariates and time-to-event. We derive motivation from the information theoretic source coding argument, and show that incorporating the knowledge of the directed acyclic graph (DAG) can be beneficial if suitable source encoders are employed. As a possible source encoder in this context, we derive a variational inference based conditional variational autoencoder for causal structured survival prediction, which we refer to as \texttt{DAGSurv}. We illustrate the performance of \texttt{DAGSurv} on low and high-dimensional synthetic datasets, and real-world datasets such as METABRIC and GBSG. We demonstrate that the proposed method outperforms other survival analysis baselines such as \texttt{Cox} Proportional Hazards, \texttt{DeepSurv} and \texttt{Deephit}, which are oblivious to the underlying causal relationship between data entities.} }
Endnote
%0 Conference Paper %T DAGSurv: Directed Ayclic Graph Based Survival Analysis Using Deep Neural Networks %A Ansh Kumar Sharma %A Rahul Kukreja %A Ranjitha Prasad %A Shilpa Rao %B Proceedings of The 13th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Vineeth N. Balasubramanian %E Ivor Tsang %F pmlr-v157-sharma21a %I PMLR %P 1065--1080 %U https://proceedings.mlr.press/v157/sharma21a.html %V 157 %X Causal structures for observational survival data provide crucial information regarding the relationships between covariates and time-to-event. We derive motivation from the information theoretic source coding argument, and show that incorporating the knowledge of the directed acyclic graph (DAG) can be beneficial if suitable source encoders are employed. As a possible source encoder in this context, we derive a variational inference based conditional variational autoencoder for causal structured survival prediction, which we refer to as \texttt{DAGSurv}. We illustrate the performance of \texttt{DAGSurv} on low and high-dimensional synthetic datasets, and real-world datasets such as METABRIC and GBSG. We demonstrate that the proposed method outperforms other survival analysis baselines such as \texttt{Cox} Proportional Hazards, \texttt{DeepSurv} and \texttt{Deephit}, which are oblivious to the underlying causal relationship between data entities.
APA
Sharma, A.K., Kukreja, R., Prasad, R. & Rao, S.. (2021). DAGSurv: Directed Ayclic Graph Based Survival Analysis Using Deep Neural Networks. Proceedings of The 13th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 157:1065-1080 Available from https://proceedings.mlr.press/v157/sharma21a.html.

Related Material