… deep joint source-channel coding. Our model incorporates the nonlinear transform as a
strong prior to effectively extract the source semantic features and provide side information for …

… , we call nonlinear transform source-channel coding (NTSCC), extract semantic latent
features of the source signal, and learn entropy model to guide joint source-channel coding with …

… In this paper, as a representative case, we inject our online adaptation method to the nonlinear
transform source-channel coding (NTSCC) based semantic communication system [24], …

… nonlinear Fourier transform at the transmitter for signals having only a discrete spectrum.
Among several methods, the Darboux transform … about sources of noise in fiber-optic channels …

… , nonlinear transforms are better suited to model the source … channel of ga and each input
channel of gs are subjected to a … affine transformation, as in (18); layer affine: each channel of …

… channel remains free from the nonlinear crosstalk that is thought to be one of the main sources
of … known to be loose for nonlinear and non-Gaussian information channels 39 , the actual …

… nonlinear mixtures known as post-nonlinear mixtures. These mixtures constituted by a linear
instantaneous mixture (linear memoryless channel… to very restrictive transforms and will not …

… mixture which is our concern, the Gaussianization transform used to compensate the
nonlinearity was proved to be effective where we can see that the result are better if the PNL …

… uses a Karhunen–Loéve transform (KLT) to decorrelate the source vector (for the … nonlinear
transformations in the analog part of the HDA system (in [6], [12], only linear transformations …

… JSCC scheme that bypasses the transformation of the pixel values to … channel input samples,
while the nonlinear activation functions allow to learn a non-linear mapping from the source …