Two ultra-dense stars spiral around each other and eventually collide. This dance lasts only moments, but it sends powerful ripples through space — gravitational waves. After the merger, a rapidly spinning blob remains, ringing like a bell, and that ring carries information about what the stars were made of.

The authors took recordings of such "rings" from computer simulations and trained neural networks to guess three key parameters: stellar mass, tidal deformability (how soft the star is, like marshmallow or hard like a nut), and the slope of the mass–radius relation (how quickly the star's size changes as mass is added). The neural network works like an experienced mechanic who can diagnose an engine by its sound. Only instead of an engine, it's a neutron star, and instead of a knock, it's the gravitational wave spectrum.

Here's the catch: this is an inverse problem. Usually, physicists go from stellar properties to predicting waves, but here it's the other way around. That's harder, because the same waves can sometimes be produced by different stars. To handle this ambiguity, the neural networks were trained in two stages: first, a rough prediction, then refinement focusing on the trickiest cases. This approach is like tuning a musical instrument: first approximate, then precise.

The results are inspiring: the neural networks don't just spit out numbers; they also reproduce known physical relationships. For example, they correctly link the frequency of the fundamental "ring" tone to the star's softness. This means the method works not as a black box but as a true physical tool. When next-generation detectors come online, such neural networks will help us peer into the interiors of neutron stars and understand what exotic forms of matter they are made of.