Inspiration

Our fascination with the universe and the idea that tiny variations in starlight could reveal new worlds inspired us to create Exo AI. We wanted to combine astronomy and AI to make exoplanet discovery faster, more accurate, and accessible to everyone.

What it does

Exo AI automatically analyzes stellar light curves to detect exoplanet transits. By combining BLS and GPI methods with machine learning, it identifies reliable candidates quickly and filters out false positives. The web platform allows both scientists and enthusiasts to explore stars, analyze data, and potentially discover new planets.

How we built it

We used Python for data processing and machine learning, C++ to optimize performance-critical algorithms, and integrated NASA TESS and Kepler datasets. Different AI models (ChatGPT, Claude, Gemini, Qwen) were used to enhance analysis and improve accuracy. The web interface was designed to make complex astronomical analysis intuitive and accessible.

Challenges we ran into

Processing millions of light curves efficiently and distinguishing weak signals from noise were major hurdles. Optimizing algorithms without losing accuracy required a lot of trial and error. Integrating multiple AI models and ensuring consistent results was also challenging.

Accomplishments that we're proud of

We developed a system that speeds up exoplanet discovery significantly, reduces human error, and empowers anyone to engage in scientific exploration. The project also successfully integrates multiple AI tools into one cohesive platform.

What we learned

We deepened our understanding of astronomical data analysis, AI integration, and software optimization. We also learned the importance of building user-friendly tools that make science accessible to both experts and the public.

What's next for Exo AI

Our plans include improving detection accuracy, expanding the dataset, adding more sophisticated machine learning models, and introducing collaborative features for citizen scientists worldwide.

Built With

Share this project:

Updates