An AI neural network has scanned 100 million archival Hubble images in just days
Sandeep K S
Jan 28
1 min read
AI Unveils Hubble's Hidden Universe: An innovative neural network, AnomalyMatch, analyzes nearly 100 million images in just 2.5 days, uncovering over 1,300 cosmic anomalies, including galaxy mergers and gravitational lenses, with many discoveries new to scientific literature.
It is one of the most famous questions in science, asked over lunch by physicist Enrico Fermi. With hundreds of billions of stars and billions of years, life should have emerged many times over. And yet, no signals. No visitors. No evidence of anyone at all.
The universe is 13 billion years old. Our galaxy alone contains hundreds of billions of stars, a significant proportion of which host planets. Many of those planets sit in the right temperature range for liquid water.
Here's a thought experiment that keeps planetary scientists awake at night. Strip every living thing from our planet—every bacterium, every blade of grass, every creature that has ever drawn breath—and ask a simple but profound question: Would Earth still be a world capable of supporting life?
The answer, it turns out, is yes. And that finding has enormous implications for how we search for life beyond our solar system. The problem is subtle but important.
Using the MeerKAT radio telescope, astronomers have discovered a natural "space laser" originating from a violently merging galaxy more than 8 billion light-years away.
When gas-rich galaxies collide, the impact compresses enormous reservoirs of gas. This violent cosmic crash can stimulate molecules—specifically hydroxyl (OH) molecules—causing them to emit incredibly bright, coherent radio waves.
Comments