Problem
Research questions and friction points this paper is trying to address.
Develop scalable deep kernel survival models
Ensure interpretability and theoretical accuracy guarantees
Handle large datasets with efficient training and testing
Survival Kernets: Scalable and Interpretable Deep Kernel Survival Analysis with an Accuracy Guarantee
🤖 AI Summary
To address the challenges of poor scalability, limited interpretability, and lack of theoretical guarantees in large-scale survival analysis, this paper proposes SurvivalKernet—the first method to introduce kernel netting into survival modeling. It compresses the training set into interpretable clusters via kernel-based clustering and enables efficient prediction using weighted cluster representations. The approach integrates XGBoost warm-starting with neural architecture search–inspired heuristic optimization to jointly enhance training efficiency and generalization. Theoretically, it establishes a finite-sample error bound with logarithmic-factor optimality. Empirically, on four standard benchmarks (up to 3 million samples), SurvivalKernet achieves significantly higher time-dependent C-indices than state-of-the-art baselines, accelerates inference by over 100×, and supports cluster-level visual interpretability.
📝 Abstract
Kernel survival analysis models estimate individual survival distributions with the help of a kernel function, which measures the similarity between any two data points. Such a kernel function can be learned using deep kernel survival models. In this paper, we present a new deep kernel survival model called a survival kernet, which scales to large datasets in a manner that is amenable to model interpretation and also theoretical analysis. Specifically, the training data are partitioned into clusters based on a recently developed training set compression scheme for classification and regression called kernel netting that we extend to the survival analysis setting. At test time, each data point is represented as a weighted combination of these clusters, and each such cluster can be visualized. For a special case of survival kernets, we establish a finite-sample error bound on predicted survival distributions that is, up to a log factor, optimal. Whereas scalability at test time is achieved using the aforementioned kernel netting compression strategy, scalability during training is achieved by a warm-start procedure based on tree ensembles such as XGBoost and a heuristic approach to accelerating neural architecture search. On four standard survival analysis datasets of varying sizes (up to roughly 3 million data points), we show that survival kernets are highly competitive compared to various baselines tested in terms of time-dependent concordance index. Our code is available at: https://github.com/georgehc/survival-kernets
Innovation
Methods, ideas, or system contributions that make the work stand out.
Deep kernel survival models
Kernel netting compression strategy
Warm-start procedure with XGBoost
🔎 Similar Papers
No similar papers found.
