“We don’t just need more observations—we also need better analysis. In fact, the planets might be there hiding in our existing dataset. If we can get the stellar activity signal removed correctly, we can uncover them.”

Yan Liang

We have more information about the cosmos than ever before, yet the search for a truly Earth-like planet around a Sun-like star remains elusive. The culprit? Stars themselves. Their turbulent surfaces distort light in ways we struggle to interpret, masking the faint gravitational tugs of orbiting worlds. For Yan Liang, the solution wasn’t just better telescopes or more data—it was generative Artificial Intelligence (AI).

Ms. Liang’s work bridges exoplanetary science and AI, harnessing machine learning to sift through astronomical data with an objective eye. Her breakthrough came when she realized that traditional methods were hitting a wall. Stellar noise was actively mimicking or erasing planetary signals. Drawing from her success in training AI to describe galaxies, Ms. Liang developed a generative model to cut through the interference and reveal what is truly there.

The result was Æstra, an AI neural network designed to decode the tangled signals of distant worlds. By training Æstra on spectral data line by line, Ms. Liang taught it to recognize the subtle distortions caused by stellar activity and distinguish them from the delicate gravitational footprint of orbiting planets. This approach allows astronomers to isolate true planetary signals without human bias. Successfully tested on simulated data, Æstra is now being validated on real observations of the Sun.

During her 51 Pegasi b Fellowship, Ms. Liang will use Æstra to comb through decades of archival data for hidden worlds. She will expand its reach to planets in the habitable zones of M dwarfs—abundant, small, cool stars—and young, highly active stars, where precise mass measurements could reshape our understanding of how planets form and evolve. Somewhere in the data, a new world awaits. With Æstra, Ms. Liang is helping bring it into focus.

Ms. Liang will receive her Ph.D. in astrophysics from Princeton University in Summer 2025.