A review of new science, in plain language

The Astrophysical Journal  ·  11 September 2026  ·  From issue 35

A Search for a More Robust Active Galactic Nucleus Classifier Using Machine Learning

New computer models find hidden black holes in galaxies, but struggle with small ones.

Alexander Messick, Vivienne Baldassare.

Small galaxies often hide black holes at their centers that older detection methods miss. Computer models trained on about 6,500 galaxies with active black holes and 53,000 without them learned to spot these hidden black holes. The models got things right 98% of the time overall. But for small, low-mass galaxies, they were only right 60% of the time.

Integrity screen: passed (3 checks) Checked 14 September 2026. Read the source

The plain-language reading of this paper is in the full issue, for subscribers.

Read the full issue

Five vetted papers in plain language, free, every two days

Five papers free every issue, about 15 emails a month. No spam. Leave in one click.

Also in issue 35 15 papers
More from The Astrophysical Journal