Saved in:
Bibliographic Details
Main Authors: Wang, Qi, Federici, Marco, van Hoof, Herke
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.03264
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915093398159360
author Wang, Qi
Federici, Marco
van Hoof, Herke
author_facet Wang, Qi
Federici, Marco
van Hoof, Herke
contents The neural process (NP) is a family of computationally efficient models for learning distributions over functions. However, it suffers from under-fitting and shows suboptimal performance in practice. Researchers have primarily focused on incorporating diverse structural inductive biases, \textit{e.g.} attention or convolution, in modeling. The topic of inference suboptimality and an analysis of the NP from the optimization objective perspective has hardly been studied in earlier work. To fix this issue, we propose a surrogate objective of the target log-likelihood of the meta dataset within the expectation maximization framework. The resulting model, referred to as the Self-normalized Importance weighted Neural Process (SI-NP), can learn a more accurate functional prior and has an improvement guarantee concerning the target log-likelihood. Experimental results show the competitive performance of SI-NP over other NPs objectives and illustrate that structural inductive biases, such as attention modules, can also augment our method to achieve SOTA performance. Our code is available at \url{https://github.com/hhq123gogogo/SI_NPs}.
format Preprint
id arxiv_https___arxiv_org_abs_2501_03264
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Bridge the Inference Gaps of Neural Processes via Expectation Maximization
Wang, Qi
Federici, Marco
van Hoof, Herke
Machine Learning
Artificial Intelligence
Neural and Evolutionary Computing
The neural process (NP) is a family of computationally efficient models for learning distributions over functions. However, it suffers from under-fitting and shows suboptimal performance in practice. Researchers have primarily focused on incorporating diverse structural inductive biases, \textit{e.g.} attention or convolution, in modeling. The topic of inference suboptimality and an analysis of the NP from the optimization objective perspective has hardly been studied in earlier work. To fix this issue, we propose a surrogate objective of the target log-likelihood of the meta dataset within the expectation maximization framework. The resulting model, referred to as the Self-normalized Importance weighted Neural Process (SI-NP), can learn a more accurate functional prior and has an improvement guarantee concerning the target log-likelihood. Experimental results show the competitive performance of SI-NP over other NPs objectives and illustrate that structural inductive biases, such as attention modules, can also augment our method to achieve SOTA performance. Our code is available at \url{https://github.com/hhq123gogogo/SI_NPs}.
title Bridge the Inference Gaps of Neural Processes via Expectation Maximization
topic Machine Learning
Artificial Intelligence
Neural and Evolutionary Computing
url https://arxiv.org/abs/2501.03264