How to Hear a Star's Age express
A star trembles like a bell after being struck. These tremors, tiny fluctuations in brightness and surface, can be caught by a telescope. From the frequencies of these 'stellar notes,' astronomers infer what's happening inside: how old the star is, how it's structured. But to decode the melody, one used to have to sift through thousands of models — slow and costly.
Now there's a tool called PITCHFORK — a neural network with a branched architecture. It learns from precomputed stellar models and then, in seconds, outputs both standard parameters (temperature, luminosity, heavy-element fraction) and the frequencies of 35 types of oscillations. Accuracy is high: temperature is predicted with an error of less than six degrees, and frequencies with a margin of just 0.02 percent.
PITCHFORK is combined with a statistical method that accounts for random errors and yields not a single number but a whole range of possible values for each parameter. It's like getting an honest 'between 4.2 and 4.8 billion years' with probabilities attached, instead of a single answer 'the star is 4.5 billion years old.'
The method was tested on the Sun and on the two components of the binary system 16 Cygni. The results matched previous reliable estimates. So the tool can be trusted. In the coming years, new space missions will bring data on millions of stars, and such a fast yet careful neural network will help us not drown in that stream and finally understand how suns evolve in different corners of the Galaxy.
Original abstract
as written by the authors arXiv:2601.02926
Accurately measuring stellar ages and internal structures is challenging, but the inclusion of asteroseismic observables can substantially improve precision. However, the curse of dimensionality means this comes at a high computational cost when using standard interpolation methods across grids of stellar models. Furthermore, without a rigorous treatment of random uncertainties in grid-based modelling, it is not possible to address systematic errors in stellar models. We present PITCHFORK -- a multilayer perceptron neural network with a branching architecture capable of rapid emulation of both classical stellar observables and individual asteroseismic oscillation modes of solar-like oscillators. PITCHFORK can predict the classical observables $$T_{\text{eff}}$$, $$L$$, and $$\left[\mathrm{Fe}/\mathrm{H}\right]$$ with precisions of $$5.88\,\text{K}$$, $$0.014\,\text{L}_{\odot}$$, and $$0.001\,\text{dex}$$, respectively, and can predict 35 individual radial mode frequencies with a uniform precision of $$0.02$$ per cent. PITCHFORK is coupled to a vectorised Bayesian inference pipeline to return well-sampled and fully marginalised posterior distributions. We validate our rigorous treatment of the random uncertainties -- including the asteroseismic surface effect -- in an extensive hare-and-hounds exercise. We also demonstrate our ability to infer the stellar properties of benchmark stars -- namely, the Sun and the binary stars 16 Cygni A and B. This work demonstrates a computationally scalable and statistically robust framework for stellar parameter inference of solar-like oscillators using individual asteroseismic mode frequencies. This provides a foundation for the treatment of systematics in preparation for the imminent abundance of asteroseismic data from future missions.
The oscillation frequencies of a star depend on its internal density: the older the star, the slower it 'rings' in certain modes, just as an aging bell changes its tone.