AI tracks plasma moving through changing black-hole jets

A new imaging method creates time-continuous videos of radio observations and estimates plasma motion between visible features.

Editorial illustration — not from the study.

Jets from supermassive black holes change over time, but existing imaging methods have had difficulty resolving that variability at high resolution. The new method represents the observations as a video and learns how the source changes across time while processing observations from different dates together.

The researchers demonstrated the approach using multi-epoch observations of the blazar 3C 345 made with the Very Long Baseline Array, a network of radio telescopes. They report improved resolution and dynamic range compared with traditional methods, allowing them to estimate the plasma's local, instantaneous velocity in the jet.

What the video reveals

The method reconstructs polarisation-sensitive, time-continuous videos from observations of sources that change quickly or slowly. In the demonstrated case, it was applied to repeated Very Long Baseline Array observations of blazar 3C 345. The resulting time-continuous reconstruction, together with improvements in resolution and dynamic range over traditional methods, enabled measurement of the local, instantaneous velocity of plasma in the jet instead of relying only on the motion of discrete components.

Why continuous motion matters

Measuring plasma motion continuously could give researchers a more complete description of how black-hole jets evolve than methods based only on identifiable knots or other separate features. The authors say the approach could be applied across monitoring programmes and to many sources, helping test or reassess models of jet motion. The abstract also describes possible use with fast-changing horizon-scale observations of Sagittarius A*, although that application is not the demonstration reported here.

Evidence and caveats

This is a journal article reporting a new reconstruction algorithm and its demonstration on multi-epoch radio observations of one blazar, 3C 345. The abstract does not give numerical measures of the resolution or dynamic-range gains, velocity estimates, or uncertainties, and it does not report validation across a larger sample. Its proposed applications to fast-changing sources and to hundreds of monitored sources remain prospective in the information provided.

A shared chemical reaction helped form both peptides and RNA-like chains

A study found that amino acids and nucleotide building blocks could jointly produce short peptides and oligonucleotides in one-pot reactions under prebiotic conditions. The linked chemistry formed a reactive intermediate that was associated with greater amounts of some oligomers than reactions using either type of building block alone.

2 hours ago·1 views

Physics & Space

Ninety-two antiprotons make a 7.5-kilometre road trip

Researchers transported 92 trapped antiprotons 7.5 kilometres by road without losing particles or measurably worsening the trap’s vacuum. The demonstration could allow future matter–antimatter tests in laboratories with less magnetic-field noise than CERN’s Antimatter Factory.

20 hours ago·5 views

Physics & Space

AI classifiers can help recreate private training images

Researchers developed an attack that uses an AI classifier to reconstruct representative images from its private training data. In tests, the approach produced more faithful images than earlier methods while maintaining similar accuracy in identifying their target classes.

1 day ago·14 views