Second-order properties of lossy likelihoods and the MLE/MDL dichotomy in lossy compression
نویسندگان
چکیده
This paper develops a theoretical framework for lossy source coding that treats it as a statistical problem, in analogy to the approach to universal lossless coding suggested by Rissanen’s Minimum Description Length (MDL) principle. Two methods for selecting efficient compression algorithms are proposed, based on lossy variants of the Maximum Likelihood and MDL principles. Their theoretical performance is analyzed, and it is shown under appropriate assumptions that the MDL approach to universal lossy coding identifies the optimal model class of lossy codes. Mokshay Madiman is with the Department of Statistics, Yale University, 24 Hillhouse Avenue, New Haven, CT 06511, USA. Email: [email protected] Ioannis Kontoyiannis is with the Division of Applied Mathematics and the Department of Computer Science, Brown University, Box F, 182 George Street, Providence, RI 02912, USA. Email: [email protected]
منابع مشابه
A Minimum Description Length Proposal for Lossy Data Compression
We give a development of the theory of lossy data compression from the point of view of statistics. This is partly motivated by the enormous success of the statistical approach in lossless compression, in particular Rissanen’s celebrated Minimum Description Length (MDL) principle. A precise characterization of the fundamental limits of compression performance is given, for arbitrary data source...
متن کاملON A LOSSY IMAGE COMPRESSION/RECONSTRUCTION METHOD BASED ON FUZZY RELATIONAL EQUATIONS
The pioneer work of image compression/reconstruction based onfuzzy relational equations (ICF) and the related works are introduced. TheICF regards an original image as a fuzzy relation by embedding the brightnesslevel into [0,1]. The compression/reconstruction of ICF correspond to thecomposition/solving inverse problem formulated on fuzzy relational equations.Optimizations of ICF can be consequ...
متن کاملModel Selection via Rate - Distortion Theory
Rissanen’s Minimum Description Length (MDL) principle for model selection proposes that, among a predetermined collection of models, we choose the one which assigns the shortest description to the data at hand. In this context, a “description” is a lossless representation of the data that also takes into account the cost of describing the chosen model itself. We examine how the MDL principle mi...
متن کاملCompact Lossy Inductance Simulators With Electronic Control
In this paper two R‑L network simulator configurations employing a single VDDIBA, one resistance and one grounded capacitance are presented. The first configuration is a grounded series resistor-inductor (R‑L) network simulator and the second configuration is intended for grounded parallel resister-inductor (R‑L) circuit simulation. Both the proposed circuits enjoy several beneficial features s...
متن کاملA Higher Order B-Splines 1-D Finite Element Analysis of Lossy Dispersive Inhomogeneous Planar Layers
In this paper we propose an accurate and fast numerical method to obtain scattering fields from lossy dispersive inhomogeneous planar layers for both TE and TM polarizations. A new method is introduced to analyze lossy Inhomogeneous Planar Layers. In this method by applying spline based Galerkin’s method of moment to scalar wave equation and imposing boundary conditions we obtain reflection and...
متن کامل