Research map / Probabilistic and scientific modeling / Molecules and proteins
Revisiting Neural Processes via Fourier Transform and Volterra Series
Peiman Mohseni, Nick Duffield, Raymond K. W. Wong
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 SFConvCNP instead of attention-based models on sawtooth | CNP family · Synthetic 1D Sawtooth | log-likelihood: -0.87 -> 1.19 | Improved |
| replace pointwise non-linearities with identity in baseline backbones (ablation) | TNP, TE-TNP, AttnCNP baselines · Synthetic 1D regression | log-likelihood | Hurt |
| remove pointwise non-linearities from SFConvCNP backbone (ablation) | SFConvCNP · Synthetic 1D regression | log-likelihood | No clear effect |
| use SFConv-based models instead of non-equivariant CNP when shifting region | ERA5 regression · ERA5 Western Europe (OOD) | log-likelihood: <-10 -> 1.71 | Improved |
| use Volterra construction (SFVConvCNP) instead of stacked SFConv blocks | SFConvCNP · Kolmogorov flow | log-likelihood: 1.89 -> 1.48 | Hurt |
| use Volterra construction (SFVConvCNP) instead of stacked SFConv blocks | SFConvCNP · ERA5 Central Europe (ID) | log-likelihood: 1.55 -> 1.09 | Hurt |
| use Volterra construction (SFVConvCNP) instead of stacked SFConv blocks | SFConvCNP · Synthetic 1D Square | log-likelihood: -0.36 -> -0.10 | Improved |
| use SFConvCNP instead of grid-based SConvCNP | Grid-SConvCNP · CIFAR-10 image completion | log-likelihood: 1.71 -> 1.74 | Improved |
Closest papers by meaning
- Spectral Convolutional Conditional Neural Process (NeurIPS 2025 · similarity 0.90)
- Translation Equivariant Transformer Neural Processes (ICML 2024 · similarity 0.89)
- Approximately Equivariant Neural Processes (NeurIPS 2024 · similarity 0.87)
- Spectral Transformer Neural Processes (NeurIPS 2026 · similarity 0.87)
- Gridded Transformer Neural Processes for Spatio-Temporal Data (ICML 2025 · similarity 0.86)
- Neural Fourier Transform: A General Approach to Equivariant Representation Learning (ICLR 2024 · similarity 0.86)
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.
