The aim of collecting data is to gain meaningful information about a phenomenon of interest. Unfortunately, often the phenomenon is not a direct physical observable. Instead, e.g., the data at hand may be a linear superposition of the desired quantities. This linear, and simplest type of information mixing is endemic in the physical sciences, arising in fields as diverse as atmospheric physics and medial diagnostics (see listing in Section 5.1). The common problem confronting workers in these fields is how to “unmix” (or, restore, enhance, de-blur, de-convolve), the data.