TreeAgent: A Generalizable Multi-Agent Framework for Automated Bias Labeling in Forestry via Compiled Expert Rules and Vision-Language Models

Shiyi Chen, Collin Hargreaves, Nicholas Saban, Huiqi Wang

Track: Track 1: Original Research/Position/Education/Attention Track

TL;DR: We present a multi-agent framework automating bias labeling in forestry by combining expert decision trees with Vision-Language Models, reducing annotation effort while outperforming ML baselines on tree height bias classification.

Abstract: Human-labeled data are widely treated as ground truth in ML, yet expert annotation is slow, inconsistent, and a major bottleneck for scaling tasks like tree height bias classification in forest remote sensing. We propose a multi-agent system (MAS) that orchestrates expert decision trees with Vision-Language Models (VLMs), treating the decision tree as a structural prior while VLMs perform localized semantic perception at individual nodes, with multi-agent voting to mitigate VLM stochasticity. We formalize a Decoupled Declarative Decision (D3) Framework that enables zero-modification generalization across diverse expert-defined decision structures. On a tree bias classification testbed, our framework outperforms supervised ML baselines and reduces human labeling effort while preserving symbolic interpretability. This suggests agentic orchestration of VLMs with expert priors is a viable path toward scalable, interpretable labeling in domains where ground truth is expensive and expert-defined.

Keywords: Multi-Agentic Systems, Automated Data Labeling, Vision-Language Models, Forestry, Remote Sensing

Submission Number: 122