Research map / Learning theory and algorithms / Prediction and decision losses
Left Barrier Loss for Unbiased Survival Analysis Prediction
Oshrit Shtossel, Bar Dubovski, Omry Koren, Yoram Louzoun
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add the left barrier RATIO loss to the MSE loss of a neural network | 2-layer FCN · SurvSet (76 datasets) | CI vs reported SurvSet results: better in 89% of datasets, average CI gain 2.5% | Improved |
| add RATIO loss to MSE instead of MSE alone or Tobit loss | 2-layer FCN · SurvSet (76 datasets) | SCC, AUC, CI: fewer than 8% of datasets showed any reduced measure | Improved |
| replace Tobit's censored-time loss with RATIO's one-sided left barrier loss | Ridge regression · Simulated survival data | error bias and variance, SCC, AUC, CI | Improved |
| increase the censoring rate | RATIO vs Ridge and Tobit · Simulated survival data | performance gain of RATIO | Improved |
| add FIESTA data augmentation from low-informative features to RATIO (ablation) | RATIO · Simulated data, fewer than 50 samples | error distribution width, CI | Improved |
| add FIESTA data augmentation to RATIO on tasks with very few uncensored samples (ablation) | RATIO · Allergy-d, Allergy-s, Mucositis-s (microbiome) | performance | Improved |
| choose Cauchy instead of exponential or hyperbolic decay for the DA weight function (ablation) | FIESTA · Simulated data | performance | No clear effect |
| iteratively regenerate censored augmented TTEs from previous augmentation (ablation) | FIESTA · Microbiome datasets | performance | No clear effect |
| use RATIO and FIESTA instead of COX, RSF, Ridge, Tobit | Microbiome survival models · Allergy and Mucositis tasks | performance | Improved |
Closest papers by meaning
- Boosting Methods for Interval-censored Data with Regression and Classification (ICLR 2025 · similarity 0.86)
- Deep Copula-Based Survival Analysis for Dependent Censoring with Identifiability Guarantees (AAAI 2024 · similarity 0.85)
- Orthogonal Survival Learners for Estimating Heterogeneous Treatment Effects from Time-to-Event Data (NeurIPS 2025 · similarity 0.85)
- History-Aware Conformal Prediction Sets for Censored Time-to-Event Outcomes (NeurIPS 2026 · similarity 0.84)
- Doubly Robust Conformalized Survival Analysis with Right-Censored Data (ICML 2025 · similarity 0.84)
- Learning Survival Models with Right-Censored Reporting Delays (NeurIPS 2026 · 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.
