AI Insight
This paper introduces SSH-Net, a deep neural network architecture designed to predict failure times when systems can fail from multiple competing causes. The model structures the neural network to align with hierarchical data organization, allowing different covariate groups to be processed through separate sub-networks before generating cause-specific hazard predictions. The approach was validated through simulation studies and applied to real-world GPU failure data from Titan supercomputers.
Why it matters
This methodology could improve reliability predictions for complex engineered systems where components can fail through different mechanisms, such as computer hardware, manufacturing equipment, or medical devices. By better capturing hierarchical data structures, the approach may provide more accurate failure forecasts, enabling better maintenance scheduling and resource allocation in industrial and research computing environments.
Understand the Science
Abstract: Competing risks are commonly observed in engineering fields and can bring challenges to time-to-event data modeling when the application scenarios are complicated. Recently, deep neural networks have received great attention for prediction with competing risks, due to their flexibility and high learning capability. However, the complexity of neural network structure brings extra difficulty in hyperparameter tuning based on different data inputs. Additionally, when an engineered system has complex physical structures with multiple hierarchical levels, treating all structural levels as a single group of inputs may fail to capture critical information. To address the issues, we propose a Structured Segmented Hazard Deep Neural Network (SSH-Net) for failure time prediction under cause-specific competing risks framework. Our approach associates neural network structure with data structures, and allows different covariate groups to impact the failure prediction through separate sub-networks. The neural network is constructed based on a cause-specific competing risks model. The SSH-Net outputs cause-specific hazard functions, and utilizes the penalized log-likelihood as the loss function. The prediction accuracy of SSH-Net is validated through simulation studies by evaluating the Brier score, the area under receiver operating characteristic curves (AUC), and the root mean square error (RMSE) of the predicted cause-specific cumulative incident function. We further demonstrate the model’s ability to predict failure time distribution functions using the Titan GPU failure time data.