Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Frozen Priors, Fluid Forecasts: Prequential Uncertainty for Low-Data Deployment with Pretrained Generative Models
Fernando Ruiz-Mazo, Vikas K Garg
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 |
|---|---|---|---|
| use DWS (Dirichlet-weighted shrinkage toward frozen GPT-2) instead of nonparametric bootstrap | uncertainty intervals · GPT-2 WikiText-2, n0=20 | coverage@90%: 0.500 -> 0.900 | Improved |
| use DWS instead of Bayesian bootstrap | uncertainty intervals · GPT-2 WikiText-2, n0=5 | coverage@90%: 0.000 -> 1.000 | Improved |
| use DWS at larger n0 | uncertainty intervals · GPT-2 WikiText-2, n0=100 | coverage@90%: 0.900 vs 0.900 (BB, JK) | No clear effect |
| use martingale posterior instead of nonparametric bootstrap | uncertainty intervals · Two Moons theta1, n0=5 | coverage@90%: 0.175 -> 0.825 | Improved |
| use martingale posterior instead of nonparametric bootstrap at large n | uncertainty intervals · Two Moons theta1, n0=1000 | coverage@90%: 0.925 -> 0.825 | Hurt |
| increase real samples n0 | MP interval · Two Moons theta1 | interval width: 0.126 -> 0.017 | Improved |
| use DWS instead of bootstrap | interval width · GPT-2 WikiText-2, small n | width | Hurt |
| use MP rather than nonparametric bootstrap on theta2 | uncertainty intervals · Two Moons theta2 | coverage@90%: n0=10: 0.975 vs 0.900; n0=20: 0.875 vs 0.900 | Mixed |
| fixed lambda (no decay) or non-predictable weights (ablation) | forecast blend · theory | scalar coherence | Hurt |
Closest papers by meaning
- Conformal Prediction Beyond the Seen: A Missing Mass Perspective for Uncertainty Quantification in Generative Models (NeurIPS 2025 · similarity 0.86)
- Probabilistic Forecasting with Stochastic Interpolants and Föllmer Processes (ICML 2024 · similarity 0.86)
- Flow Matching with Gaussian Process Priors for Probabilistic Time Series Forecasting (ICLR 2025 · similarity 0.85)
- Quantifying Uncertainty in the Presence of Distribution Shifts (NeurIPS 2025 · similarity 0.85)
- ProbTS: Benchmarking Point and Distributional Forecasting across Diverse Prediction Horizons (NeurIPS 2024 · similarity 0.85)
- ABC: Any-Subset Autoregression via Non-Markovian Diffusion Bridges in Continuous Time and Space (NeurIPS 2026 · similarity 0.85)
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.
