Nonlinear inverse problems are ubiquitous in medical sciences, engineering, physics and other natural sciences. This workshop will advance a statistical paradigm that allows underpinning statistical decision making in such problems in a scientifically rigorous way. This will be achieved by providing `frequentist' large sample guarantees for inferences arising from commonly used (Bayesian or non-Bayesian) algorithms in such problems. Thereby, a large cluster of applied problems, where algorithms are used for day-to-day decision making with statistical data, will be put on a solid and robust foundation, facilitating its confident use in modern society.