Rehan Sameer

تخصص الذكاء الاصطناعي · دفعة 2026

LacteaChat: A Natural Language Interface For Exploring 2 Billion Stars from the Gaia Catalog

The Gaia mission is transforming astronomy by constructing a 3D map of nearly two billion stars in the Milky Way and beyond. Despite this unprecedented dataset, the scientific value of stellar spectra remains underutilized due to the lack of large-scale interactive systems for spectral exploration. We present LacteaChat, a project that extends the open-source Gaia spectral visualization system with a conversational AI interface for semantic exploration of stellar spectra and astronomical structures. The system allows users to interact with the visualization pipeline using natural language commands such as navigating to regions of interest, filtering stars by physical properties, recoloring the scene according to spectral characteristics, highlighting unusual objects, and querying astrophysical populations such as white dwarfs or cool stars. Unlike conventional image based AI systems, the proposed framework operates directly on physically meaningful spectral representations and metadata generated by the rendering pipeline, including per-pixel accumulated spectra and stellar attributes. The project investigates how large language modelscan serve asa coordination layer for scientific visualization by translating user intent into structured visualization, filtering, and analysis operations while relying on deterministic GPU-based backends to preserve scientific correctness. The resulting system aims to reduce the barrier to exploration of billion-scale astronomical datasets and enable more intuitive, interactive, and discovery-driven analysis workflows accessible to all.

اقرأ التقرير النهائي (PDF)