Saved in:
Bibliographic Details
Main Authors: Lara-Rangel, Jose, Chen, Nanze, Zhang, Fengzhe
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.14416
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916698783744000
author Lara-Rangel, Jose
Chen, Nanze
Zhang, Fengzhe
author_facet Lara-Rangel, Jose
Chen, Nanze
Zhang, Fengzhe
contents Neural Processes (NPs) have gained attention in meta-learning for their ability to quantify uncertainty, together with their rapid prediction and adaptability. However, traditional NPs are prone to underfitting. Transformer Neural Processes (TNPs) significantly outperform existing NPs, yet their applicability in real-world scenarios is hindered by their quadratic computational complexity relative to both context and target data points. To address this, pseudo-token-based TNPs (PT-TNPs) have emerged as a novel NPs subset that condense context data into latent vectors or pseudo-tokens, reducing computational demands. We introduce the Induced Set Attentive Neural Processes (ISANPs), employing Induced Set Attention and an innovative query phase to improve querying efficiency. Our evaluations show that ISANPs perform competitively with TNPs and often surpass state-of-the-art models in 1D regression, image completion, contextual bandits, and Bayesian optimization. Crucially, ISANPs offer a tunable balance between performance and computational complexity, which scale well to larger datasets where TNPs face limitations.
format Preprint
id arxiv_https___arxiv_org_abs_2504_14416
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Exploring Pseudo-Token Approaches in Transformer Neural Processes
Lara-Rangel, Jose
Chen, Nanze
Zhang, Fengzhe
Machine Learning
Neural Processes (NPs) have gained attention in meta-learning for their ability to quantify uncertainty, together with their rapid prediction and adaptability. However, traditional NPs are prone to underfitting. Transformer Neural Processes (TNPs) significantly outperform existing NPs, yet their applicability in real-world scenarios is hindered by their quadratic computational complexity relative to both context and target data points. To address this, pseudo-token-based TNPs (PT-TNPs) have emerged as a novel NPs subset that condense context data into latent vectors or pseudo-tokens, reducing computational demands. We introduce the Induced Set Attentive Neural Processes (ISANPs), employing Induced Set Attention and an innovative query phase to improve querying efficiency. Our evaluations show that ISANPs perform competitively with TNPs and often surpass state-of-the-art models in 1D regression, image completion, contextual bandits, and Bayesian optimization. Crucially, ISANPs offer a tunable balance between performance and computational complexity, which scale well to larger datasets where TNPs face limitations.
title Exploring Pseudo-Token Approaches in Transformer Neural Processes
topic Machine Learning
url https://arxiv.org/abs/2504.14416