What frustrates professional astrophysicists about current exoplanet detection algorithms?
I'm part of a student team working on an AI-based exoplanet transit detection pipeline using TESS light curves for a hackathon.
We're building a system that detects periodic dips, classifies them (transit, eclipsing binary, false positive, noise, blends, etc.), and estimates parameters like period, depth, and duration.
I'd love to hear from people who have worked with TESS, Kepler, exoplanet surveys, or astronomical time-series data:
- What are the biggest pain points with current transit detection pipelines?
- What types of false positives cause the most trouble?
- What information do existing tools fail to provide clearly?
- What would make candidate vetting faster or more reliable?
- Are there common mistakes student projects make when working with light curves?
- If you could add one feature to existing exoplanet detection software, what would it be?
I'm especially interested in issues around noisy data, crowded fields, blending, uncertainty estimation, explainability, and candidate ranking.