Orion-MSP: Multi-Scale Sparse Attention for Tabular In-Context Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bouadi, Mohamed, Seth, Pratinav, Tanna, Aditya, Sankarapu, Vinay Kumar
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915604711079936
author Bouadi, Mohamed
Seth, Pratinav
Tanna, Aditya
Sankarapu, Vinay Kumar
author_facet Bouadi, Mohamed
Seth, Pratinav
Tanna, Aditya
Sankarapu, Vinay Kumar
contents Tabular data remain the predominant format for real-world applications. Yet, developing effective neural models for tabular data remains challenging due to heterogeneous feature types and complex interactions occurring at multiple scales. Recent advances in tabular in-context learning (ICL), such as TabPFN and TabICL, have achieved state-of-the-art performance comparable to gradient-boosted trees (GBTs) without task-specific fine-tuning. However, current architectures exhibit key limitations: (1) single-scale feature processing that overlooks hierarchical dependencies, (2) dense attention with quadratic scaling in table width, and (3) strictly sequential component processing that prevents iterative representation refinement and cross-component communication. To address these challenges, we introduce Orion-MSP, a tabular ICL architecture featuring three key innovations: (1) multi-scale processing to capture hierarchical feature interactions; (2) block-sparse attention combining windowed, global, and random patterns for scalable efficiency and long-range connectivity; and (3) a Perceiver-style memory enabling safe bidirectional information flow across components. Across diverse benchmarks, Orion-MSP matches or surpasses state-of-the-art performance while scaling effectively to high-dimensional tables, establishing a new standard for efficient tabular in-context learning. The model is publicly available at https://github.com/Lexsi-Labs/Orion-MSP .
format Preprint
id arxiv_https___arxiv_org_abs_2511_02818
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Orion-MSP: Multi-Scale Sparse Attention for Tabular In-Context Learning
Bouadi, Mohamed
Seth, Pratinav
Tanna, Aditya
Sankarapu, Vinay Kumar
Artificial Intelligence
Machine Learning
Tabular data remain the predominant format for real-world applications. Yet, developing effective neural models for tabular data remains challenging due to heterogeneous feature types and complex interactions occurring at multiple scales. Recent advances in tabular in-context learning (ICL), such as TabPFN and TabICL, have achieved state-of-the-art performance comparable to gradient-boosted trees (GBTs) without task-specific fine-tuning. However, current architectures exhibit key limitations: (1) single-scale feature processing that overlooks hierarchical dependencies, (2) dense attention with quadratic scaling in table width, and (3) strictly sequential component processing that prevents iterative representation refinement and cross-component communication. To address these challenges, we introduce Orion-MSP, a tabular ICL architecture featuring three key innovations: (1) multi-scale processing to capture hierarchical feature interactions; (2) block-sparse attention combining windowed, global, and random patterns for scalable efficiency and long-range connectivity; and (3) a Perceiver-style memory enabling safe bidirectional information flow across components. Across diverse benchmarks, Orion-MSP matches or surpasses state-of-the-art performance while scaling effectively to high-dimensional tables, establishing a new standard for efficient tabular in-context learning. The model is publicly available at https://github.com/Lexsi-Labs/Orion-MSP .
title Orion-MSP: Multi-Scale Sparse Attention for Tabular In-Context Learning
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2511.02818