When compressive learning fails: blame the decoder or the sketch?
This addresses a theoretical problem for researchers in compressive learning, but appears incremental as it focuses on analyzing existing methods.
The paper investigates why compressive learning fails by analyzing the non-convex optimization landscape and heuristics like CLOMPR through numerical simulations, but does not report specific results or numbers.
In compressive learning, a mixture model (a set of centroids or a Gaussian mixture) is learned from a sketch vector, that serves as a highly compressed representation of the dataset. This requires solving a non-convex optimization problem, hence in practice approximate heuristics (such as CLOMPR) are used. In this work we explore, by numerical simulations, properties of this non-convex optimization landscape and those heuristics.