Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Beyond Unimodal: Generalising Neural Processes for Multimodal Uncertainty Estimation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 multimodal Bayesian aggregation rather than concatenating modality representations (ablation) | MNPs · CIFAR10-C | test ECE: 0.125±0.005 -> 0.011±0.001 | Helped |
| Use multimodal Bayesian aggregation rather than averaging modalities (ablation) | MNPs · CIFAR10-C | OOD AUC (SVHN): 0.803±0.007 -> 0.872±0.002 | Helped |
| Use multimodal Bayesian aggregation rather than averaging modality representations (ablation) | MNPs · HMDB | test accuracy: 57.80±1.97 -> 71.97±0.43 | Helped |
| Use class-error-based MSE context updates instead of random sampling (ablation) | MNPs · CUB | test ECE: 0.069±0.029 -> 0.049±0.008 | Helped |
| Use MSE context selection rather than CE context selection (ablation) | MNPs · CIFAR10-C | OOD AUC (CIFAR100): 0.789±0.004 -> 0.786±0.005 | Mixed |
| Select new context points using classification error rather than random sampling (ablation) | MNPs · CIFAR10-C | test ECE: 0.073±0.005 -> 0.011±0.001 | Helped |
| Use adaptive RBF attention with its supervised contrastive loss and lengthscale penalty (ablation) | MNPs · all evaluated datasets | test accuracy | Helped |
| Use Sparsemax rather than Softmax with adaptive RBF attention (ablation) | MNPs · all evaluated datasets | test accuracy | Helped |
| Increase the number of context points (ablation) | MNPs · robustness and OOD datasets | test accuracy | Helped |
| Increase the number of context points (ablation) | MNPs · robustness and OOD datasets | ECE | Mixed |
| Use MNPs rather than MGP | multimodal uncertainty estimation · PIE | testing speed-up: 5.37 | Helped |
| Use MNPs rather than MGP | multimodal uncertainty estimation · CIFAR10-C with SVHN OOD | OOD AUC: 0.803±0.007 -> 0.872±0.002 | Helped |
Closest papers by meaning
- Distance-informed Neural Processes (NeurIPS 2025 · similarity 0.87)
- Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance Awareness (NeurIPS 2020 · similarity 0.86)
- Rényi Neural Processes (ICML 2025 · similarity 0.86)
- Test Time Scaling for Neural Processes (NeurIPS 2025 · similarity 0.86)
- Uncertainty Quantification with the Empirical Neural Tangent Kernel (NeurIPS 2025 · similarity 0.84)
- One Step Closer to Unbiased Aleatoric Uncertainty Estimation (AAAI 2024 · 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.