Decoding cosmic signals with deep learning. Building neural networks for exoplanet detection, lunar phase classification, and celestial data exploration.
A deep dive into exoplanet detection, transit photometry, and the journey from a simple CNN model to state-of-the-art multi-class vetting architectures for Kepler and TESS satellite data.
2D CNN • CLASSIFIER
Deep learning lunar phase recognition with custom dataset augmentation, 4-phase classification, and live interactive scanning simulation.
NLP • PERSPECTIVE AI
Enter the view you think is correct, explore counter-perspectives, and challenge bias with intelligent natural language processing algorithms.
TIME SERIES • ASTRONOMY
Stellar light curve time-series processing, phase folding, BLS & TLS periodogram pipelines on Kepler and TESS NASA datasets.
A 2-month deep dive born from an ISRO hackathon problem statement. Exploring transit photometry, Kepler light curves, and the evolution of machine learning architectures from Kepler Autovetter and AstroNet to 5-class vetting models and ExoMiner++.
✦ 10-Part Scientific Breakdown
I'm Debomoy Patra, a Machine Learning developer and space tech enthusiast deeply passionate about astroinformatics. I love building end-to-end architectures that translate complex celestial light signals into actionable discoveries.
From manually crafting and augmenting lunar phase datasets for Moon CNN, to analyzing Kepler light curves and exoplanetary transit dynamics for Astro Py — I believe every dataset holds an undiscovered story waiting for the right neural network.