IEEE Open Journal of Signal Processing

The fundamental task of classification given a limited number of training data samples is considered for physicalsystems with known parametric statistical models. The standalone learning-based and statistical model-based classifiers face major challenges towards the fulfillment of the classification task using a small training set. Specifically, classifiers that solely rely on the physics-based statistical models usually suffer from their inability to properly tune the underlying unobservable parameters, which leads to a mismatched representation of the system’s behaviors.

We develop a novel 2D functional learning framework that employs a sparsity-promoting regularization based on second-order derivatives. Motivated by the nature of the regularizer, we restrict the search space to the span of piecewise-linear box splines shifted on a 2D lattice. Our formulation of the infinite-dimensional problem on this search space allows us to recast it exactly as a finite-dimensional one that can be solved using standard methods in convex optimization.

The paper develops novel algorithms for time-varying (TV) sparse channel estimation in Massive multiple-input, multiple-output (MMIMO) systems. This is achieved by employing a novel reduced (non-uniformly spaced tap) delay-line equalizer, which can be related to low/reduced rank filters. This low rank filter is implemented by deriving an innovative TV (Krylov-space based) Multi-Stage Kalman Filter (MSKF), employing appropriate state estimation techniques.

The end users’ satisfactory Quality of Experience (QoE) is a fundamental criterion for networked video service providers such as video-on-demand providers (Netflix, YouTube, etc.), cloud gaming providers (Google Stadia, PlayStation Now, etc.) and videoconferencing providers (Zoom, Microsoft Teams, etc.). To know the QoE, providers today typically predict it from the Quality of Service (QoS) parameters or the client-side's actual QoE metrics measured at the current time-step.

In the era of big data, profitable opportunities are becoming available for many applications. As the amount of data keeps increasing, machine learning becomes an attractive tool to analyze the information acquired. However, harnessing meaningful data remains a challenge. The machine learning tools employed in many applications apply all training data without taking into consideration how relevant are some of them. In this paper, we propose a data selection strategy for the training step of Neural Networks to obtain the most significant data information and improve algorithm performance during training. 

Target source extractionis significant for improving human speech intelligibility and the speech recognition performance of computers. This study describes a method for target source extraction, called the similarity-and-independence-awarebeamformer (SIBF). The SIBF extracts the target source using a rough magnitude spectrogram as the reference signal. The advantage of the SIBF is that it can obtain a more accurate signal than the spectrogram generated by target-enhancing methods such as speech enhancement based on deep neural networks. 

Quadrature spatial modulation (QSM) isa recently proposed multiple-input multiple-output (MIMO) wireless transmission paradigm that has garnered considerable research interest owing to its relatively high spectral efficiency. QSM essentially enhances the spatial multiplexing gain while maintaining all the inherent advantages of spatial modulation (SM).

Identification of decompressed JPEG images, especially those compressed with high JPEG quality factors, is a challenging issue in image forensics. Furthermore, the applicability of the existing JPEG forensic detectors in forgery localization is limited by their inability to cope with spatial misalignment in the 8×8 JPEG grid.

This work exploits Riemannian manifolds to build a sequential-clustering framework able to address a wide variety of clustering tasks in dynamic multilayer (brain) networks via the information extracted from their nodal time-series. The discussion follows a bottom-up path, starting from feature extraction from time-series and reaching up to Riemannian manifolds (feature spaces) to address clustering tasks such as state clustering, community detection (a.k.a. network-topology identification), and subnetwork-sequence tracking. 

We present a structured overview of adaptation algorithms for neural network-based speech recognition, considering both hybrid hidden Markov model / neural network systems and end-to-end neural network systems, with a focus on speaker adaptation, domain adaptation, and accent adaptation.