A Pattern Language for Resilient Visual Agents

• URL: http://arxiv.org/abs/2604.28001v1
• Date: Thu, 30 Apr 2026 15:24:26 GMT
• Title: A Pattern Language for Resilient Visual Agents
• Abstract summary: We propose an architectural pattern language for visual agents that separates fast, deterministic reflexes from slow, probabilistic supervision.<n>It consists of four architectural design patterns: (1) Hybrid Affordance Integration, (2) Adaptive Visual Anchoring, (3) Visual Hierarchy Synthesis, and (4) Semantic Scene Graph.
• Score: 69.08074737002393
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: 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.

Related papers

• A Two-Dimensional Framework for AI Agent Design Patterns: Cognitive Function and Execution Topology [54.72352855252792]
  We propose a two-dimensional classification that combines a Cognitive Function axis with seven categories and an Execution Topology axis with six structural archetypes.<n>The resulting 7x6 matrix identifies 27 named patterns, 13 with original names and provides a principled, framework-neutral, and model-agnostic vocabulary for AI agent architecture design.
  arXiv  Detail & Related papers  (2026-03-16T04:01:01Z)
• UVLM: A Universal Vision-Language Model Loader for Reproducible Multimodal Benchmarking [0.42970700836450476]
  This paper introduces UVLM, a Google Colab-based framework that provides a unified interface for loading,numeric, and benchmarking VLM architectures on custom image analysis tasks.<n>UVLM supports two major model families -- LLaVA-NeXT and Qwen2.5-VL -- which differ fundamentally in their vision encoding, tokenization, and decoding strategies.<n>Key features include a multi-task prompt builder with support for four response types, a consensus validation mechanism based on majority voting, a flexible token budget, and a built-in chain-of-thought reference mode for benchmarking.
  arXiv  Detail & Related papers  (2026-03-14T11:07:42Z)
• Agentic Design Patterns: A System-Theoretic Framework [8.108572809924956]
  Existing efforts to agentic design patterns often lack a rigorous systems-theoretic foundation.<n>We propose a novel system-theoretic framework that deconstructs an agentic AI system into five core, interacting functional subsystems.<n>We present a collection of 12 agentic design patterns that offer reusable, structural solutions to recurring problems in agent design.
  arXiv  Detail & Related papers  (2026-01-27T16:14:08Z)
• Uni4D-LLM: A Unified SpatioTemporal-Aware VLM for 4D Understanding and Generation [61.60600246983274]
  Existing 3D and 4D approaches typically embed scene geometry into autogressive model for semantic understanding and diffusion model for content generation.<n>We propose Uni4D-LLM, the first unified VLM framework withtemporal awareness for 4D scene understanding and generation.
  arXiv  Detail & Related papers  (2025-09-28T12:06:54Z)
• When Better Eyes Lead to Blindness: A Diagnostic Study of the Information Bottleneck in CNN-LSTM Image Captioning Models [0.0]
  This paper documents a systematic, iterative development of foundational image captioning models.<n>It progresses from a simple CNN-LSTM encoder-decoder to a competitive attention-based system.<n>The experiments chart the impact of architectural enhancements and demonstrate a key finding within the classic CNN-LSTM paradigm.
  arXiv  Detail & Related papers  (2025-07-24T20:20:44Z)
• Modeling and Visualization Reasoning for Stakeholders in Education and Industry Integration Systems: Research on Structured Synthetic Dialogue Data Generation Based on NIST Standards [3.5516803380598074]
  This study addresses the structural complexity and semantic ambiguity in stakeholder interactions within the Education-Industry Integration (EII) system.<n>We propose a structural modeling paradigm based on the National Institute of Standards and Technology (NIST) synthetic data quality framework.
  arXiv  Detail & Related papers  (2025-06-20T12:37:43Z)
• Topological Structure Learning Should Be A Research Priority for LLM-Based Multi-Agent Systems [69.95482609893236]
  Large Language Model-based Multi-Agent Systems (MASs) have emerged as a powerful paradigm for tackling complex tasks through collaborative intelligence.<n>We call for a paradigm shift toward emphtopology-aware MASs that explicitly model and dynamically optimize the structure of inter-agent interactions.
  arXiv  Detail & Related papers  (2025-05-28T15:20:09Z)
• LLM-based Hierarchical Concept Decomposition for Interpretable Fine-Grained Image Classification [5.8754760054410955]
  We introduce textttHi-CoDecomposition, a novel framework designed to enhance model interpretability through structured concept analysis. Our approach not only aligns with the performance of state-of-the-art models but also advances transparency by providing clear insights into the decision-making process.
  arXiv  Detail & Related papers  (2024-05-29T00:36:56Z)
• NPHardEval4V: Dynamic Evaluation of Large Vision-Language Models with Effects of Vision [64.83085920775316]
  We introduce NPHardEval4V, a multimodal benchmark suite grounded in four classical NP-hard problems.<n>Each task is presented through a combination of structured visual layouts and textual prompts, designed to assess the ability of LVLMs to perform reasoning under visual-linguistic constraints.<n>Our results show that while these models perform reasonably well on perception-based inputs, they struggle with global optimization, abstraction, and constraint satisfaction.
  arXiv  Detail & Related papers  (2024-03-04T07:10:31Z)
• Prismatic VLMs: Investigating the Design Space of Visually-Conditioned Language Models [73.40350756742231]
  Visually-conditioned language models (VLMs) have seen growing adoption in applications such as visual dialogue, scene understanding, and robotic task planning. Despite the volume of new releases, key design decisions around image preprocessing, architecture, and optimization are under-explored.
  arXiv  Detail & Related papers  (2024-02-12T18:21:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.