Astronomical surveys & machine learning – Ting-Yun Cheng
Existing astronomical datasets have reached a historic high, providing far more data than can be analysed within limited timeframes. The imminent dawn of next-generation surveys will further generate unprecedented data volumes, fundamentally challenging every aspect of how we observe, store, process, and analyse astronomical data. Managing this information tsunami requires a definitive shift away from traditional analysis toward automated, scalable solutions. In this talk, I will introduce how machine learning (ML) techniques have become vital tools in modern astronomy, driving critical efficiencies in data processing and catalogue construction. Beyond these pipeline improvements, I will also introduce how ML methodologies are being used to explore new scientific insights, with the extraction of subtle physical properties from complex datasets, bridging between simulation and observation, and introducing new perspectives from computer vision.
