Neural Fields Reconstruct VLBI Videos for Plasma Velocity Measurement

A novel algorithm named kine neural reconstruction has been developed to produce high-resolution, time-continuous very long baseline interferometry (VLBI) videos. This advancement, detailed in a publication in Nature on August 26, 2026, with the digital object identifier 10.1038/s41586-026-10988-5, allows for the direct measurement of instantaneous plasma velocities. This capability is crucial for understanding dynamic astrophysical phenomena.

Very Long Baseline Interferometry (VLBI) is a technique that synchronizes multiple radio telescopes to create a virtual telescope with a diameter equivalent to the distance between the farthest telescopes. This allows astronomers to achieve extremely high angular resolution, enabling them to observe distant and faint celestial objects. Traditionally, VLBI data processing has focused on producing static images of astronomical sources. However, many astrophysical processes, such as the ejection of material from black holes and active galactic nuclei, are inherently dynamic and occur over very short timescales. Capturing these rapid changes requires the ability to reconstruct time-varying information from VLBI observations.

The kine neural reconstruction algorithm addresses this challenge by employing neural fields, a class of deep learning models that can represent complex, continuous functions. By training these neural fields on VLBI data, the algorithm can learn to generate a sequence of images that represent the evolution of the observed source over time. This results in videos that are not only high-resolution but also time-continuous, meaning there are no abrupt jumps or missing frames between observations. This temporal continuity is essential for accurately tracking the movement and changes within astrophysical structures.

The primary benefit of this new method is the direct measurement of instantaneous plasma velocities. Previously, estimating plasma velocities from VLBI data often involved inferring motion from sequences of static images, which could be imprecise and limited by the cadence of observations. The kine neural reconstruction algorithm, by providing a continuous temporal representation, allows for the calculation of velocities at any given moment. This opens up new avenues for detailed kinematic analysis of relativistic astrophysical jets, which are powerful outflows of plasma ejected from the vicinity of supermassive black holes and other compact objects. Understanding the dynamics of these jets is fundamental to comprehending phenomena such as gamma-ray bursts and the evolution of galaxies. The algorithm's ability to provide detailed kinematic analysis means researchers can now study the intricate movements, acceleration, and deceleration of plasma within these jets with unprecedented accuracy, potentially leading to breakthroughs in our understanding of high-energy astrophysics and the behavior of matter under extreme conditions.