Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Bayesian Transfer Learning for Artificially Intelligent Geospatial Systems: A Predictive Stacking Approach
Luca Presicce, Sudipto Banerjee
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 disagreement tempering (ablation) | DBPS (K=5) · Synthetic spatial GP (50 replicates) | predictive interval width: 2.03 -> 1.96 | Improved |
| add disagreement tempering (ablation) | DBPS (K=10) · Synthetic spatial GP (50 replicates) | predictive interval width: 2.06 -> 1.99 | Improved |
| add disagreement tempering (ablation) | DBPS (K=20) · Synthetic spatial GP (50 replicates) | predictive interval width: 2.19 -> 2.04 | Improved |
| add disagreement tempering (ablation) | DBPS (K=20) · Synthetic spatial GP (50 replicates) | empirical coverage: 0.960 -> 0.947 | Hurt |
| add disagreement tempering (ablation) | DBPS (K=5) · Synthetic spatial GP (50 replicates) | empirical coverage: 0.946 -> 0.939 | Hurt |
| add disagreement tempering (ablation) | DBPS (K=10) · Synthetic spatial GP (50 replicates) | MSPE: 0.268 -> 0.267 | No clear effect |
| increase number of subsets K from 5 to 20 | DBPS · Synthetic spatial GP (50 replicates) | computation time (sec): 4.8 -> 1.6 | Improved |
| increase number of subsets K from 5 to 20 | DBPS · Synthetic spatial GP (50 replicates) | predictive interval width: 2.03 -> 2.19 | Hurt |
| increase number of subsets K from 5 to 20 | DBPS · Synthetic spatial GP (50 replicates) | MSPE: 0.265 -> 0.271 | Hurt |
| increase number of subsets K from 5 to 20 | DBPS · Synthetic spatial GP (50 replicates) | empirical coverage: 0.946 -> 0.960 | Mixed |
| replace full GP with DBPS (K=5) | Full GP · Synthetic spatial GP (50 replicates) | computation time (sec): 1654 -> 4.8 | Improved |
| replace full GP with NNGP (m=5) | Full GP · Synthetic spatial GP (50 replicates) | computation time (sec): 1654 -> 17 | Improved |
Closest papers by meaning
- Deep Random Features for Scalable Interpolation of Spatiotemporal Data (ICLR 2025 · similarity 0.86)
- Spatial meshing for general Bayesian multivariate models (JMLR 2024 · similarity 0.86)
- Causal Deciphering and Inpainting in Spatio-Temporal Dynamics via Diffusion Model (NeurIPS 2024 · similarity 0.85)
- Spatial Multivariate Trees for Big Data Bayesian Regression (JMLR 2022 · similarity 0.85)
- Drawing Informative Gradients from Sources: A One-stage Transfer Learning Framework for Cross-city Spatiotemporal Forecasting (AAAI 2025 · similarity 0.85)
- Unifying Knowledge from Diverse Datasets to Enhance Spatial-Temporal Modeling: A Granularity-Adaptive Geographical Embedding Approach (ICML 2025 · 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.
