Signal separation in the passive underwater acoustic domain has heavily relied on deep learning techniques to isolate ship radiated noise. However, the separation networks commonly used in this domain stem from speech separation applications and may not fully consider the unique aspects of underwater acoustics beforehand, such as the influence of different propagation media, signal frequencies and modulation characteristics. This oversight highlights the need for tailored approaches that account for the specific characteristics of underwater sound propagation. This study introduces a novel temporal network designed to separate ship radiated noise by employing a dual-path model and a feature decoupling approach. The mixed signals' features are transformed into a space where they exhibit greater independence, with each dimension's significance decoupled. Subsequently, a fusion of local and global attention mechanisms is employed in the separation layer. Extensive comparisons showcase the effectiveness of this method when compared to other prevalent network models, as evidenced by its performance in the ShipsEar and DeepShip datasets.

Blind signal separation (BSS) is an important and challenging signal processing task. Given an observed signal which is a superposition of a collection of unknown (hidden/latent) signals, BSS aims at recovering the separate, underlying signals from only the observed mixed signal. As an underdetermined problem, BSS is notoriously difficult to solve in general, and modern deep learning has provided engineers with an effective set of tools to solve this problem. For example, autoencoders learn a low-dimensional hidden encoding of the input data which can then be used to perform signal separation. In real-time systems, a common bottleneck is the transmission of data (communications) to a central command in order to await decisions. Bandwidth limits dictate the frequency and resolution of the data being transmitted. To overcome this, compressed sensing (CS) technology allows for the direct acquisition of compressed data with a near optimal reconstruction guarantee. This paper addresses the question: can compressive acquisition be combined with deep learning for BSS to provide a complete acquire-separate-predict pipeline? In other words, the aim is to perform BSS on a compressively acquired signal directly without ever having to decompress the signal. We consider image data (MNIST and E-MNIST) and show how our compressive autoencoder approach solves the problem of compressive BSS. We also provide some theoretical insights into the problem.

ABSTRACT In this paper, we address the intricate issue of RF signal separation by presenting a novel adaptation of the WaveNet architecture that introduces learnable dilation parameters, significantly enhancing signal separation in dense RF spectrums. Our focused architectural refinements and innovative data augmentation strategies have markedly improved the model's ability to discern complex signal sources. This paper details our comprehensive methodology, including the refined model architecture, data preparation techniques, and the strategic training strategy that have been pivotal to our success. The efficacy of our approach is evidenced by the substantial improvements recorded: a 58.82% increase in SINR at a BER of 10 Notably, our model achieved first place in the challenge [1], demonstrating its Figure 1: Modified Wavenet with Learnable Dilation and superior performance and establishing a new standard for Padding machine learning applications within the RF communications domain. Index Terms-- Radio Frequency Signal Separation, Machine Learning, WaveNet Architecture, Learnable Dilation, Data Augmentation 1. INTRODUCTION The co-channel signal separation in the crowded radiofrequency Figure 1: An Illustration of Learnable Dilation Rate (RF) spectrum is a crucial task for enabling various wireless systems to operate simultaneously.

Due to the influences of ocean environment noise and sea water channels, the separation of underwater acoustic signals is a challenging problem. Some studies have researched the separation of underwater signals by separating the different components of signals with different characteristics, such as spatial orientation information and category differences, in a certain signal transformation domain. Some methods separate signals directly on the feature domain based on expert knowledge [1-3]. The wrap transform was used to separate dispersive time-frequency components in [1]. A depth-based method was proposed in [2], where the modified Fourier transformation of the output power of a plane-wave beamformer was used to separate the signals obtained from a vertical line array. In [3], rigid and elastic acoustic scattering components of underwater target echoes were separated in the fractional Fourier transform domain based on a target echo highlight model. Most other algorithms rely on blind signal separation (BSS) methods [4-10]. In [4], the frequency components of the Detection of Envelope Modulation on Noise (DEMON) spectrum were used to separate signals in different directions via independent component analysis (ICA). According to the main frequency bands of different signals in a linear superposition signal, in [5], bandpass filters were used first, and then eigenvalue decomposition was employed for separation purposes [6] and [7] used the Sawada algorithm and ideal binary masking to separate artificially mixed whale songs.

Single-channel signal separation and deconvolution aims to separate and deconvolve individual sources from a single-channel mixture and is a challenging problem in which no prior knowledge of the mixing filters is available. Both individual sources and mixing filters need to be estimated. In addition, a mixture may contain non-stationary noise which is unseen in the training set. We propose a synthesizing-decomposition (S-D) approach to solve the single-channel separation and deconvolution problem. In synthesizing, a generative model for sources is built using a generative adversarial network (GAN). In decomposition, both mixing filters and sources are optimized to minimize the reconstruction error of the mixture. The proposed S-D approach achieves a peak-to-noise-ratio (PSNR) of 18.9 dB and 15.4 dB in image inpainting and completion, outperforming a baseline convolutional neural network PSNR of 15.3 dB and 12.2 dB, respectively and achieves a PSNR of 13.2 dB in source separation together with deconvolution, outperforming a convolutive non-negative matrix factorization (NMF) baseline of 10.1 dB.

This paper introduces a cross adversarial source separation (CASS) framework via autoencoder, a new model that aims at separating an input signal consisting of a mixture of multiple components into individual components defined via adversarial learning and autoencoder fitting. CASS unifies popular generative networks like auto-encoders (AEs) and generative adversarial networks (GANs) in a single framework. The basic building block that filters the input signal and reconstructs the $i$-th target component is a pair of deep neural networks $\mathcal{EN}_i$ and $\mathcal{DE}_i$ as an encoder for dimension reduction and a decoder for component reconstruction, respectively. The decoder $\mathcal{DE}_i$ as a generator is enhanced by a discriminator network $\mathcal{D}_i$ that favors signal structures of the $i$-th component in the $i$-th given dataset as guidance through adversarial learning. In contrast with existing practices in AEs which trains each Auto-Encoder independently, or in GANs that share the same generator, we introduce cross adversarial training that emphasizes adversarial relation between any arbitrary network pairs $(\mathcal{DE}_i,\mathcal{D}_j)$, achieving state-of-the-art performance especially when target components share similar data structures.

We study the problem of semi-supervised singing voice separation, in which the training data contains a set of samples of mixed music (singing and instrumental) and an unmatched set of instrumental music. Our solution employs a single mapping function g, which, applied to a mixed sample, recovers the underlying instrumental music, and, applied to an instrumental sample, returns the same sample. The network g is trained using purely instrumental samples, as well as on synthetic mixed samples that are created by mixing reconstructed singing voices with random instrumental samples. Our results indicate that we are on a par with or better than fully supervised methods, which are also provided with training samples of unmixed singing voices, and are better than other recent semi-supervised methods.

ABSTRACT Source separation problems are ubiquitous in the physical sciences; any situation where signals are superimposed calls for source separation to estimate the original signals. In this tutorial I will discuss the Bayesian approach to the source separation problem. This approach has a specific advantage in that it requires the designer to explicitly describe the signal model in addition to any other information or assumptions that go into the problem description. This leads naturally to the idea of informed source separation, where the algorithm design incorporates relevant information about the specific problem. This approach promises to enable researchers to design their own high-quality algorithms that are specifically tailored to the problem at hand. 1. UNDERSTANDING THE PROBLEM To gather information about the physical world, we deploy sensors to make measurements and detect signals. Our sensors, if properly designed, will collect information about the signals of interest. However, very often the signals of interest are comprised of a set of discrete signals, which have been superimposed during propagation, often with signals that are not of interest. Thus our sensors almost invariably detect a mixture of signals--some interesting and some noninteresting.

Two well known classes of unsupervised procedures that can be cast in this manner are generative and recoding models. In a generative unsupervised framework, the environment generates training exampleswhich we will refer to as observations-by sampling from one distribution; the other distribution is embodied in the model. Examples of generative frameworks are mixtures of Gaussians (MoG) [2], factor analysis [4], and Boltzmann machines [8]. In the recoding unsupervised framework, the model transforms points from an obser- vation space to an output space, and the output distribution is compared either to a reference distribution or to a distribution derived from the output distribution. An example is independent component analysis (leA) [11], a method that discovers a representation of vector-valued observations in which the statistical dependence among the vector elements in the output space is minimized.

Two well known classes of unsupervised procedures that can be cast in this manner are generative and recoding models. In a generative unsupervised framework, the environment generates training exampleswhich we will refer to as observations-by sampling from one distribution; the other distribution is embodied in the model. Examples of generative frameworks are mixtures of Gaussians (MoG) [2], factor analysis [4], and Boltzmann machines [8]. In the recoding unsupervised framework, the model transforms points from an obser- vation space to an output space, and the output distribution is compared either to a reference distribution or to a distribution derived from the output distribution. An example is independent component analysis (leA) [11], a method that discovers a representation of vector-valued observations in which the statistical dependence among the vector elements in the output space is minimized.