Evaluating and Enhancing Negation Comprehension in Remote Sensing MLLMs

2026-06-18

AI Summary

This study addresses a critical limitation in multimodal large language models (MLLMs) for remote sensing: their poor comprehension of negation semantics, which leads to frequent hallucinations and an inability to recognize missing or incorrect content in images, thereby hindering deployment in high-stakes scenarios such as emergency response. To systematically investigate this issue, the authors introduce RS-Neg, the first benchmark specifically designed for evaluating negation understanding in remote sensing. They further propose NeFo, a test-time learning approach that leverages a large language model to generate diverse negation queries and integrates a dynamic visual grounding verification module with explicit negation logic modeling. Remarkably, NeFo achieves substantial improvements in generalization using only 5% of unlabeled test samples. Experimental results demonstrate that NeFo effectively mitigates hallucination and consistently outperforms existing methods on both seen and unseen tasks.

Abstract

Multimodal Large Language Models (MLLMs) have demonstrated remarkable success in various Remote Sensing (RS) tasks. However, their ability to comprehend negation remains underexplored, limiting deployment in real-world applications where models must explicitly identify what is false or absent, e.g., emergency responders need to locate non-flooded routes for evacuation. To comprehensively study this limitation, we introduce RS-Neg, the first benchmark to evaluate negation understanding across region-level to scene-level tasks. Specifically, we design an automated data generation pipeline for RS imagery, using LLMs to synthesize diverse negation queries, and introduce a dynamic visual focus module for verification. Our evaluation reveals that advanced RS MLLMs struggle with negation, exhibiting hallucinations and substantial performance degradation. To close this gap, we propose NeFo, a novel test-time learning method that explicitly incorporates the logical role of negation into the model optimization. Remarkably, using about 5\% unlabeled test samples, NeFo significantly improves the negation understanding of models and shows strong generalization to unseen tasks. Code and data will be released upon acceptance.

Problem

• negation comprehension
• Remote Sensing
• Multimodal Large Language Models
• hallucination
• absence reasoning

Innovation

• negation comprehension
• Remote Sensing MLLMs
• test-time learning
• RS-Neg benchmark
• NeFo