Gaussian Approximation of Quantization Error for Estimation From Compressed Data
نویسندگان
چکیده
We consider the distributional connection between lossy compressed representation of a high-dimensional signal X using random spherical code and observation under an additive white Gaussian noise (AWGN). show that Wasserstein distance bitrate- R version its AWGN-channel signal-to-noise ratio 2 2R -1 is bounded in problem dimension. utilize this fact to connect risk estimator based on attained by same when fed AWGN-corrupted X. demonstrate usefulness deriving various novel results for inference problems compression constraints, including minimax estimation, sparse regression, sensing, universality linear estimation remote source coding.
منابع مشابه
Excess Error , Approximation Error , and Estimation Error
Another question of interest concerns the behaviour of a learning algorithm in the infinite sample limit: as it receives more and more data, does the algorithm converge to an optimal prediction rule, i.e. does the generalization error of the learned function approach the optimal error? Recall that for a distribution D on X × Y and a loss ` : Y × Y→[0,∞), the optimal error w.r.t. D and ` is the ...
متن کامل3 Approximation Error vs . Estimation Error
A loss function is a mapping l : Y×Y "→ R (sometimes R×R "→ R). For example, in binary classification the 0/1 loss function l(y, p) = I(y ≠ p) is often used and in regression the squared error loss function l(y, p) = (y − p) is often used. Other loss functions include the following: absolute loss, Huber loss, εinsensitive loss, hinge loss, logistic loss, exponential loss, modified least squares...
متن کاملQuantization Distortion in Block Transform-Compressed Data
Block transform-compression techniques such as JPEG operate by systematically decomposing image data into segments or blocks that are transformed, quantized, and encoded independently. Characteristic quantization noise patterns that include artiicial discontinu-ities between data blocks result from such image compression schemes. Here we present a simple model of the quantization noise that res...
متن کاملEstimation of Primary Quantization Matrix in Double Compressed JPEG Images
In this report, we present a method for estimation of primary quantization matrix from a double compressed JPEG image. We first identify characteristic features that occur in DCT histograms of individual coefficients due to double compression. Then, we present 3 different approaches that estimate the original quantization matrix from double compressed images. Finally, most successful of them Ne...
متن کاملVariational Gaussian approximation for Poisson data
The Poisson model is frequently employed to describe count data, but in a Bayesian context it leads to an analytically intractable posterior probability distribution. In this work, we analyze a variational Gaussian approximation to the posterior distribution arising from the Poisson model with a Gaussian prior. This is achieved by seeking an optimal Gaussian distribution minimizing the Kullback...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE Transactions on Information Theory
سال: 2021
ISSN: ['0018-9448', '1557-9654']
DOI: https://doi.org/10.1109/tit.2021.3083271