Research map / Probabilistic and scientific modeling / Molecules and proteins
Adversarial Time-to-Event Modeling
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use adversarial DATE rather than parametric log-normal AFT DRAFT | DATE · EHR non-censored test set | median relative absolute error: 36.7 -> 23.6 | Helped |
| use adversarial DATE rather than parametric log-normal AFT DRAFT | DATE · flchain non-censored test set | median relative absolute error: 26.2 -> 19.5 | Helped |
| use adversarial DATE rather than parametric log-normal AFT DRAFT | DATE · support non-censored test set | median relative absolute error: 2.0 -> 2.7 | Hurt |
| use adversarial DATE rather than parametric log-normal AFT DRAFT | DATE · seer non-censored test set | median relative absolute error: 23.7 -> 18.6 | Helped |
| use an adversarial autoencoder for covariates rather than imputing missing covariates (ablation) | DATE · EHR, flchain, support, and seer non-censored test sets | median relative absolute error: EHR 23.6 -> 24.5; flchain 19.5 -> 19.3; support 2.7 -> 1.5; seer 18.6 -> 20.2 | Mixed |
| add stochastic noise to all generator layers instead of using input-layer noise only (ablation) | DATE · support non-censored test-set distributions | median 95% interval width: 22.4 -> 149.9 | Helped |
| add stochastic noise to all generator layers instead of using input-layer noise only (ablation) | DATE · support censored test-set distributions | median 95% interval width: 86.9 -> 1397.3 | Helped |
| use Uniform(0,1) noise rather than Uniform(-1,1) noise on all generator layers (ablation) | DATE · support non-censored test-set distributions | median 95% interval width: 60.0 -> 149.9 | Helped |
| use Uniform(0,1) noise rather than Gaussian(0,1) noise on all generator layers (ablation) | DATE · support non-censored test-set distributions | median 95% interval width: 37.9 -> 149.9 | Helped |
| add an ordering cost function to the DATE objective (ablation) | DATE · time-to-event ordering | concordance index | No clear effect |
| extend DRAFT to a mixture of log-normal distributions with different variances but shared mean (ablation) | DRAFT · time-to-event prediction | relative absolute error | No clear effect |
Closest papers by meaning
- Gradient-based Explanations for Deep Learning Survival Models (ICML 2025 · similarity 0.87)
- Orthogonal Survival Learners for Estimating Heterogeneous Treatment Effects from Time-to-Event Data (NeurIPS 2025 · similarity 0.85)
- A Scalable Nonparametric Continuous-Time Survival Model through Numerical Quadrature (NeurIPS 2026 · similarity 0.84)
- XTSFormer: Cross-Temporal-Scale Transformer for Irregular-Time Event Prediction in Clinical Applications (AAAI 2025 · similarity 0.84)
- SurvivalPFN: Amortizing Survival Prediction via In-Context Bayesian Inference (NeurIPS 2026 · similarity 0.84)
- MOTOR: A Time-to-Event Foundation Model For Structured Medical Records (ICLR 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.