Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Estimating Uncertainty Intervals from Collaborating Networks
Tianhui Zhou, Yitong Li, Yuan Wu, David Carlson
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 |
|---|---|---|---|
| learn f jointly (collaborating networks) instead of fixing f to uniform (ablation) | CN-g · 1-d synthetic sin example | convergence of gof with sample size: U-g and CN-f converge much slower than CN-g, which has a small margin to T-g | Improved |
| use theoretically optimal f (T-g) instead of learned f (ablation) | g network · 1-d synthetic sin example | interval sharpness | Improved |
| use full CN instead of fixed uniform f (U-g) (ablation) | g network · 1-d synthetic sin example | median and interval estimates | Improved |
| add CN-g over MSE or quantile regression | CN · 1-d synthetic overparameterized network | overfitting | Improved |
| use CN-g instead of DP on Weibull data | CN-g · Synthetic Weibull | gof: -2.384 -> -1.843 | Improved |
| use CN-g instead of DP on Weibull data | CN-g · Synthetic Weibull | calibration error cal (%): 22.424 -> 2.158 | Improved |
| use CN-g instead of EN on Gaussian data | CN-g · Synthetic heteroskedastic Gaussian | gof: -1.684 -> -1.684 | No clear effect |
| use CN-g instead of CQR on Gaussian data | CN-g · Synthetic heteroskedastic Gaussian | MAE: 0.728 -> 0.064 | Improved |
| add joint f learning (CN-g vs g-only) on Weibull (ablation) | CN-g · Synthetic Weibull | MAE: 0.053 -> 0.036 | Improved |
| add post-hoc recalibration CR to DP (ablation) | DP · Airline | calibration error cal (%): 16.16 -> 0.64 | Improved |
| add joint f learning (CN-g vs g-only) on small datasets (ablation) | CN-g · CPU | gof: -1.031 -> -1.053 | Hurt |
| use CN-f alone instead of CN-g (ablation) | CN-f · Weibull | gof: -1.843 -> -2.066 | Hurt |
Closest papers by meaning
- Distribution-Free Data Uncertainty for Neural Network Regression (ICLR 2025 · similarity 0.86)
- Uncertainty Quantification for Deep Regression using Contextualised Normalizing Flows (NeurIPS 2025 · similarity 0.84)
- Fully Heteroscedastic Count Regression with Deep Double Poisson Networks (ICML 2025 · similarity 0.84)
- Quantifying Uncertainty in the Presence of Distribution Shifts (NeurIPS 2025 · similarity 0.83)
- Adaptive Sampling to Reduce Epistemic Uncertainty Using Prediction Interval-Generation Neural Networks (AAAI 2025 · similarity 0.83)
- Boosting Methods for Interval-censored Data with Regression and Classification (ICLR 2025 · similarity 0.83)
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.
