Seeing Clearly without Training: Mitigating Hallucinations in Multimodal LLMs for Remote Sensing
This addresses hallucinations in remote sensing AI, which is an incremental improvement for domain-specific applications.
The paper tackles hallucinations in multimodal large language models for remote sensing visual question-answering by introducing a benchmark for diagnosis and a training-free inference method, resulting in improved performance and reduced hallucinations across diverse models.
Multimodal large language models (MLLMs) suffer from pronounced hallucinations in remote sensing visual question-answering (RS-VQA), primarily caused by visual grounding failures in large-scale scenes or misinterpretation of fine-grained small targets. To systematically analyze these issues, we introduce RSHBench, a protocol-based benchmark for fine-grained diagnosis of factual and logical hallucinations. To mitigate grounding-induced factual hallucinations, we further propose Relative Attention-Driven Actively Reasoning (RADAR), a training-free inference method that leverages intrinsic attention in MLLMs to guide progressive localization and fine-grained local reasoning at test time. Extensive experiments across diverse MLLMs demonstrate that RADAR consistently improves RS-VQA performance and reduces both factual and logical hallucinations. Code and data will be publicly available at: https://github.com/MiliLab/RADAR