Saved in:
Bibliographic Details
Main Authors: Furuya, Takashi, Murari, Davide, Schönlieb, Carola-Bibiane
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.15503
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911451985215488
author Furuya, Takashi
Murari, Davide
Schönlieb, Carola-Bibiane
author_facet Furuya, Takashi
Murari, Davide
Schönlieb, Carola-Bibiane
contents Stability and robustness are critical for deploying Transformers in safety-sensitive settings. A principled way to enforce such behavior is to constrain the model's Lipschitz constant. However, approximation-theoretic guarantees for architectures that explicitly preserve Lipschitz continuity have yet to be established. In this work, we bridge this gap by introducing a class of gradient-descent-type in-context Transformers that are Lipschitz-continuous by construction. We realize both MLP and attention blocks as explicit Euler steps of negative gradient flows, ensuring inherent stability without sacrificing expressivity. We prove a universal approximation theorem for this class within a Lipschitz-constrained function space. Crucially, our analysis adopts a measure-theoretic formalism, interpreting Transformers as operators on probability measures, to yield approximation guarantees independent of token count. These results provide a rigorous theoretical foundation for the design of robust, Lipschitz continuous Transformer architectures.
format Preprint
id arxiv_https___arxiv_org_abs_2602_15503
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Approximation Theory for Lipschitz Continuous Transformers
Furuya, Takashi
Murari, Davide
Schönlieb, Carola-Bibiane
Machine Learning
Stability and robustness are critical for deploying Transformers in safety-sensitive settings. A principled way to enforce such behavior is to constrain the model's Lipschitz constant. However, approximation-theoretic guarantees for architectures that explicitly preserve Lipschitz continuity have yet to be established. In this work, we bridge this gap by introducing a class of gradient-descent-type in-context Transformers that are Lipschitz-continuous by construction. We realize both MLP and attention blocks as explicit Euler steps of negative gradient flows, ensuring inherent stability without sacrificing expressivity. We prove a universal approximation theorem for this class within a Lipschitz-constrained function space. Crucially, our analysis adopts a measure-theoretic formalism, interpreting Transformers as operators on probability measures, to yield approximation guarantees independent of token count. These results provide a rigorous theoretical foundation for the design of robust, Lipschitz continuous Transformer architectures.
title Approximation Theory for Lipschitz Continuous Transformers
topic Machine Learning
url https://arxiv.org/abs/2602.15503