Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Multi-task Learning for Aggregated Data using Gaussian Processes
Fariba Yousefi, Michael T Smith, Mauricio Álvarez
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 the second task's larger-support observations to predict the first task in its test gap | Multi-task GP · Synthetic data test gap | SMSE: 0.9699± 0.016 -> 0.464 ± 0.136 | Improved |
| Use the second task's larger-support observations to predict the first task in its test gap | Multi-task GP · Synthetic data test gap | SNLP: -0.095± 0.036 -> -0.822 ± 0.109 | Improved |
| Add high-frequency, low-accuracy OPC observations to the low-resolution data | Air pollution multi-task GP · Air pollution prediction, 9 days | SMSE: 0.657±0.100 -> 0.439±0.114 | Improved |
| Use aggregated-task data alongside sparse original-resolution data | Fertility dataset multi-task GP · Canadian fertility rates, 5×5 and 2×2 aggregation | SMSE | Improved |
| Increase original-resolution training observations to 500 or more | Fertility dataset multi-task GP · Canadian fertility rates | SMSE | No clear effect |
| Use 2×2 rather than 5×5 aggregated support | Fertility dataset multi-task GP · Canadian fertility rates | SMSE | Improved |
| Model the output variance as a latent function as well as the mean (ablation) | Fertility dataset HetGPA · Canadian fertility rates, four outputs | SNLP | Improved |
Closest papers by meaning
- Variational Learning on Aggregate Outputs with Gaussian Processes (NeurIPS 2018 · similarity 0.86)
- Multi-task Sparse Structure Learning with Gaussian Copula Models (JMLR 2016 · similarity 0.86)
- Efficient Modeling of Latent Information in Supervised Learning using Gaussian Processes (NeurIPS 2017 · similarity 0.85)
- Multi-Task Processes (ICLR 2022 · similarity 0.85)
- Analysing Multi-Task Regression via Random Matrix Theory with Application to Time Series Forecasting (NeurIPS 2024 · similarity 0.85)
- Learning from Summarized Data: Gaussian Process Regression with Sample Quasi-Likelihood (AAAI 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.
