Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

Test Time Scaling for Neural Processes

NeurIPS 2025 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Priors and functional variational uncertainty · 0 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

The paper applies test-time scaling ideas to Neural Processes to improve uncertainty estimation under limited supervision, motivated by poorly fitted variational posteriors for the global latent variable.

Problem

Uncertainty-aware meta-learning with Neural Processes, where variational posteriors for the global latent variable are reported (in the abstract, truncated) to be problematic, hurting reliable uncertainty estimation under limited supervision.

Approach

Applies test-time scaling to Neural Processes so that the global latent variable posterior is refined at inference time rather than relying solely on the training-time variational posterior; the abstract is truncated so the exact mechanism is not stated.

Surprising findings

• Empirically, variational posteriors for the global latent variable in NPs are reported to be poorly fitted (abstract truncated).

Experimental setup

Builds on: Neural Processes

Closest papers by meaning

• Transformer Neural Processes: Uncertainty-Aware Meta Learning Via Sequence Modeling (ICML 2022 · similarity 0.88)
• Distance-informed Neural Processes (NeurIPS 2025 · similarity 0.87)
• Quantifying Uncertainty in the Presence of Distribution Shifts (NeurIPS 2025 · similarity 0.87)
• Rényi Neural Processes (ICML 2025 · similarity 0.85)
• Uncertainty Quantification with the Empirical Neural Tangent Kernel (NeurIPS 2025 · similarity 0.85)
• Dimension Agnostic Neural Processes (ICLR 2025 · similarity 0.85)