Designing Domain-Specific Agents via Hierarchical Task Abstraction Mechanism

2025-11-21

AI Summary

General-purpose LLM agents exhibit limited performance in specialized domains like remote sensing, primarily due to their inability to model structured workflows—e.g., multi-stage radiometric correction, spectral index computation, and interdependent intermediate artifacts. Method: We propose EarthAgent, a hierarchical multi-agent architecture guided by task dependency graphs, which replaces role-based simulation with domain-aligned, layered planning and collaborative execution. It integrates LLM-driven toolchain invocation and an interpretable task decomposition mechanism. We further introduce GeoPlan-bench—the first benchmark for geospatial analysis—featuring three evaluation dimensions: tool selection accuracy, path similarity, and logical completeness. Contribution/Results: Experiments demonstrate that EarthAgent significantly outperforms state-of-the-art single- and multi-agent systems on GeoPlan-bench, validating the effectiveness and advancement of aligning domain-specific task structures with agent architecture design.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Multiagent Systems: Multiagent Planning
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work

Abstract

LLM-driven agents, particularly those using general frameworks like ReAct or human-inspired role-playing, often struggle in specialized domains that necessitate rigorously structured workflows. Fields such as remote sensing, requiring specialized tools (e.g., correction, spectral indices calculation), and multi-step procedures (e.g., numerous intermediate products and optional steps), significantly challenge generalized approaches. To address this gap, we introduce a novel agent design framework centered on a Hierarchical Task Abstraction Mechanism (HTAM). Specifically, HTAM moves beyond emulating social roles, instead structuring multi-agent systems into a logical hierarchy that mirrors the intrinsic task-dependency graph of a given domain. This task-centric architecture thus enforces procedural correctness and decomposes complex problems into sequential layers, where each layer's sub-agents operate on the outputs of the preceding layers. We instantiate this framework as EarthAgent, a multi-agent system tailored for complex geospatial analysis. To evaluate such complex planning capabilities, we build GeoPlan-bench, a comprehensive benchmark of realistic, multi-step geospatial planning tasks. It is accompanied by a suite of carefully designed metrics to evaluate tool selection, path similarity, and logical completeness. Experiments show that EarthAgent substantially outperforms a range of established single- and multi-agent systems. Our work demonstrates that aligning agent architecture with a domain's intrinsic task structure is a critical step toward building robust and reliable specialized autonomous systems.

Problem

• Addresses LLM agents' struggles in specialized domains requiring structured workflows
• Proposes hierarchical task abstraction to enforce procedural correctness in complex tasks
• Focuses on domain-specific multi-agent systems for reliable autonomous problem-solving

Innovation

• Hierarchical Task Abstraction Mechanism structures multi-agent systems
• Task-centric architecture ensures procedural correctness and decomposition
• EarthAgent instantiates framework for complex geospatial analysis