Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
Research into a technique called inertial confinement fusion aims to enhance nuclear-fusion performance in laboratory experiments. Improvements in the technique have been made using a clever statistical approach.
The pursuit of thermonuclear fusion, the power source of stars, in the laboratory is an ambitious endeavour. For a useful number of fusions to occur, fusion fuel must be heated to tens of millions of degrees so that it produces an ionized gas called a plasma. If such a plasma could be confined for long enough, the energy released by fusions, known as the yield, would greatly exceed the energy invested in the plasma — a long-elusive goal of fusion researchers. In inertial confinement fusion (ICF) experiments, the fusion plasma is generated when high-power drivers, such as lasers, are used to implode fusion fuel. Writing in Nature, Gopalaswamy et al.1 report the use of experimentally trained statistical models to triple the fusion yield and substantially improve the plasma confinement in ICF experiments.