A Pattern Language for Resilient Visual Agents

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gidey, Habtom Kahsay, Lenz, Alexander, Knoll, Alois
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909004910821376
author Gidey, Habtom Kahsay
Lenz, Alexander
Knoll, Alois
author_facet Gidey, Habtom Kahsay
Lenz, Alexander
Knoll, Alois
contents Integrating multimodal foundation models into enterprise ecosystems presents a fundamental software architecture challenge. Architects must balance competing quality attributes: the high latency and non-determinism of vision language action (VLA) models versus the strict determinism and real-time performance required by enterprise control loops. In this study, we propose an architectural pattern language for visual agents that separates fast, deterministic reflexes from slow, probabilistic supervision. It consists of four architectural design patterns: (1) Hybrid Affordance Integration, (2) Adaptive Visual Anchoring, (3) Visual Hierarchy Synthesis, and (4) Semantic Scene Graph.
format Preprint
id arxiv_https___arxiv_org_abs_2604_28001
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A Pattern Language for Resilient Visual Agents
Gidey, Habtom Kahsay
Lenz, Alexander
Knoll, Alois
Artificial Intelligence
Software Engineering
D.2.11; I.2.0
Integrating multimodal foundation models into enterprise ecosystems presents a fundamental software architecture challenge. Architects must balance competing quality attributes: the high latency and non-determinism of vision language action (VLA) models versus the strict determinism and real-time performance required by enterprise control loops. In this study, we propose an architectural pattern language for visual agents that separates fast, deterministic reflexes from slow, probabilistic supervision. It consists of four architectural design patterns: (1) Hybrid Affordance Integration, (2) Adaptive Visual Anchoring, (3) Visual Hierarchy Synthesis, and (4) Semantic Scene Graph.
title A Pattern Language for Resilient Visual Agents
topic Artificial Intelligence
Software Engineering
D.2.11; I.2.0
url https://arxiv.org/abs/2604.28001