Reproduce “Towards Faithful Reasoning in Remote Sensing: A Perceptually-Grounded GeoSpatial Chain-of-Thought for Vision-Language Models” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Jiaqi Liu, Lang Sun, Ronghao Fu, Bo Yang

RSThinker uses perceptually grounded Geospatial Chain-of-Thought to answer remote-sensing questions by planning what to inspect, linking observations to image locations, and synthesizing an answer. The authors train it with rationale supervision followed by reinforcement learning using task-metric rewards. On HRRSD counting, its mean absolute error was 0.242 versus 0.787 for ChatGPT-5. Traces expose errors but do not guarantee faithful reasoning.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.