
A problem of determining the causes (e.g., interactions) from desired effects (e.g., a pattern).An inverse problem is a class of problems where the goal is to determine the underlying causes or parameters of a system from observed effects or measurements. In physics, this often means reconstructing the properties of an object or interaction from the data it produces. For example, in the provided research, scientists use the inverse problem to infer the mass hierarchy of neutrinos from the signals of a supernova explosion, or to map dark matter distribution from gravitational lensing observations.
The formal concept of inverse problems was established in the early 20th century, with significant contributions from Hadamard (1902) who defined well-posedness, and later by Tikhonov (1963) who introduced regularization methods for unstable solutions.
The forward problem is well-defined: given a set of parameters, compute the expected observations. The inverse problem reverses this, but it is often ill-posed — multiple parameter sets can produce the same observations, and small measurement errors can lead to wildly different solutions. To solve it, physicists use statistical methods like Bayesian inference, regularization techniques, or iterative algorithms that compare simulated predictions with actual data, refining the parameters until a best fit is achieved.
Inverse problems are used in medical imaging (CT scans and MRI), geophysics (seismic tomography), and astrophysics (reconstructing the properties of distant objects from telescope data).
The inverse problem is so fundamental that even the human brain solves it constantly — interpreting the 2D image on the retina to infer the 3D structure of the world.



