Skip to yearly menu bar Skip to main content


Poster

High Dimensional Structured Superposition Models

Qilong Gu · Arindam Banerjee

Area 5+6+7+8 #94

Keywords: [ (Other) Statistics ] [ Learning Theory ] [ Sparsity and Feature Selection ] [ Model Selection and Structure Learning ] [ Regularization and Large Margin Methods ]


Abstract:

High dimensional superposition models characterize observations using parameters which can be written as a sum of multiple component parameters, each with its own structure, e.g., sum of low rank and sparse matrices. In this paper, we consider general superposition models which allow sum of any number of component parameters, and each component structure can be characterized by any norm. We present a simple estimator for such models, give a geometric condition under which the components can be accurately estimated, characterize sample complexity of the estimator, and give non-asymptotic bounds on the componentwise estimation error. We use tools from empirical processes and generic chaining for the statistical analysis, and our results, which substantially generalize prior work on superposition models, are in terms of Gaussian widths of suitable spherical caps.

Live content is unavailable. Log in and register to view live content