In-context Learning of Evolving Data Streams with Tabular Foundational Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lourenço, Afonso, Gama, João, Xing, Eric P., Marreiros, Goreti
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908710004064256
author Lourenço, Afonso
Gama, João
Xing, Eric P.
Marreiros, Goreti
author_facet Lourenço, Afonso
Gama, João
Xing, Eric P.
Marreiros, Goreti
contents State-of-the-art data stream mining has long drawn from ensembles of the Very Fast Decision Tree, a seminal algorithm honored with the 2015 KDD Test-of-Time Award. However, the emergence of large tabular models, i.e., transformers designed for structured numerical data, marks a significant paradigm shift. These models move beyond traditional weight updates, instead employing in-context learning through prompt tuning. By using on-the-fly sketches to summarize unbounded streaming data, one can feed this information into a pre-trained model for efficient processing. This work bridges advancements from both areas, highlighting how transformers' implicit meta-learning abilities, pre-training on drifting natural data, and reliance on context optimization directly address the core challenges of adaptive learning in dynamic environments. Exploring real-time model adaptation, this research demonstrates that TabPFN, coupled with a simple sliding memory strategy, consistently outperforms ensembles of Hoeffding trees, such as Adaptive Random Forest, and Streaming Random Patches, across all non-stationary benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2502_16840
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle In-context Learning of Evolving Data Streams with Tabular Foundational Models
Lourenço, Afonso
Gama, João
Xing, Eric P.
Marreiros, Goreti
Machine Learning
Artificial Intelligence
State-of-the-art data stream mining has long drawn from ensembles of the Very Fast Decision Tree, a seminal algorithm honored with the 2015 KDD Test-of-Time Award. However, the emergence of large tabular models, i.e., transformers designed for structured numerical data, marks a significant paradigm shift. These models move beyond traditional weight updates, instead employing in-context learning through prompt tuning. By using on-the-fly sketches to summarize unbounded streaming data, one can feed this information into a pre-trained model for efficient processing. This work bridges advancements from both areas, highlighting how transformers' implicit meta-learning abilities, pre-training on drifting natural data, and reliance on context optimization directly address the core challenges of adaptive learning in dynamic environments. Exploring real-time model adaptation, this research demonstrates that TabPFN, coupled with a simple sliding memory strategy, consistently outperforms ensembles of Hoeffding trees, such as Adaptive Random Forest, and Streaming Random Patches, across all non-stationary benchmarks.
title In-context Learning of Evolving Data Streams with Tabular Foundational Models
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2502.16840