top of page



Decoding the Void: How AI is Solving Space Mysteries Humans Can't See
๐ The Scene For ten years, the Kepler telescope stared at a patch of darkness, recording light from 150,000 stars. It generated terabytes of dataโmessy, noisy static. Human astronomers analyzed the brightest signals, finding thousands of planets. But they missed the faint ones. Then, a neural network trained by Google and NASA was unleashed on the "rejected" data. Within hours, it spotted a tiny, weak dip in brightness around a star called Kepler-90. It was an eighth planet,
ย
ย
ย
bottom of page
