PyTorchKR Papers

최신 논문

총 205편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2204.01355

Target Confusion in End-to-end Speaker Extraction: Analysis and Approaches

Recently, end-to-end speaker extraction has attracted increasing attention and shown promising results. However, its performance is often inferior to that of a blind source separation (BSS) counterpart with a similar network architecture, due to the auxiliary speaker encoder may sometimes generate ambiguous speaker embeddings. Such ambiguous guidance information may confuse the separation network and hence lead to wrong extraction results, which deteriorates the overall performance. We refer to this as the target confusion problem. In this paper, we conduct an analysis of such an issue and solve it in two stages. In the training phase, we propose to integrate metric learning methods to improve the distinguishability of embeddings produced by the speaker encoder. While for inference, a novel post-filtering strategy is designed to revise the wrong results. Specifically, we first identify these confusion samples by measuring the similarities between output estimates and enrollment utterances, after which the true target sources are recovered by a subtraction operation. Experiments show that performance improvement of more than 1dB SI-SDRi can be brought, which validates the effectiveness of our methods and emphasizes the impact of the target confusion problem.

arXiv:2204.00210

Spatial Loss for Unsupervised Multi-channel Source Separation

We propose a spatial loss for unsupervised multi-channel source separation. The proposed loss exploits the duality of direction of arrival (DOA) and beamforming: the steering and beamforming vectors should be aligned for the target source, but orthogonal for interfering ones. The spatial loss encourages consistency between the mixing and demixing systems from a classic DOA estimator and a neural separator, respectively. With the proposed loss, we train the neural separators based on minimum variance distortionless response (MVDR) beamforming and independent vector analysis (IVA). We also investigate the effectiveness of combining our spatial loss and a signal loss, which uses the outputs of blind source separation as the reference. We evaluate our proposed method on synthetic and recorded (LibriCSS) mixtures. We find that the spatial loss is most effective to train IVA-based separators. For the neural MVDR beamformer, it performs best when combined with a signal loss. On synthetic mixtures, the proposed unsupervised loss leads to the same performance as a supervised loss in terms of word error rate. On LibriCSS, we obtain close to state-of-the-art performance without any labeled training data.

arXiv:2203.14324

Blind Source Separation for Mixture of Sinusoids with Near-Linear Computational Complexity

We propose a multi-tone decomposition algorithm that can find the frequencies, amplitudes and phases of the fundamental sinusoids in a noisy observation sequence. Under independent identically distributed Gaussian noise, our method utilizes a maximum likelihood approach to estimate the relevant tone parameters from the contaminated observations. When estimating $M$ number of sinusoidal sources, our algorithm successively estimates their frequencies and jointly optimizes their amplitudes and phases. Our method can also be implemented as a blind source separator in the absence of the information about $M$. The computational complexity of our algorithm is near-linear, i.e., $\tilde{O}(N)$.

arXiv:2203.06306

DURRNet: Deep Unfolded Single Image Reflection Removal Network

Single image reflection removal problem aims to divide a reflection-contaminated image into a transmission image and a reflection image. It is a canonical blind source separation problem and is highly ill-posed. In this paper, we present a novel deep architecture called deep unfolded single image reflection removal network (DURRNet) which makes an attempt to combine the best features from model-based and learning-based paradigms and therefore leads to a more interpretable deep architecture. Specifically, we first propose a model-based optimization with transform-based exclusion prior and then design an iterative algorithm with simple closed-form solutions for solving each sub-problems. With the deep unrolling technique, we build the DURRNet with ProxNets to model natural image priors and ProxInvNets which are constructed with invertible networks to impose the exclusion prior. Comprehensive experimental results on commonly used datasets demonstrate that the proposed DURRNet achieves state-of-the-art results both visually and quantitatively.

arXiv:2203.02941

Single microphone speaker extraction using unified time-frequency Siamese-Unet

In this paper we present a unified time-frequency method for speaker extraction in clean and noisy conditions. Given a mixed signal, along with a reference signal, the common approaches for extracting the desired speaker are either applied in the time-domain or in the frequency-domain. In our approach, we propose a Siamese-Unet architecture that uses both representations. The Siamese encoders are applied in the frequency-domain to infer the embedding of the noisy and reference spectra, respectively. The concatenated representations are then fed into the decoder to estimate the real and imaginary components of the desired speaker, which are then inverse-transformed to the time-domain. The model is trained with the Scale-Invariant Signal-to-Distortion Ratio (SI-SDR) loss to exploit the time-domain information. The time-domain loss is also regularized with frequency-domain loss to preserve the speech patterns. Experimental results demonstrate that the unified approach is not only very easy to train, but also provides superior results as compared with state-of-the-art (SOTA) Blind Source Separation (BSS) methods, as well as commonly used speaker extraction approach.

arXiv:2202.04648

A survey of unsupervised learning methods for high-dimensional uncertainty quantification in black-box-type problems

Constructing surrogate models for uncertainty quantification (UQ) on complex partial differential equations (PDEs) having inherently high-dimensional $\mathcal{O}(10^{\ge 2})$ stochastic inputs (e.g., forcing terms, boundary conditions, initial conditions) poses tremendous challenges. The curse of dimensionality can be addressed with suitable unsupervised learning techniques used as a pre-processing tool to encode inputs onto lower-dimensional subspaces while retaining its structural information and meaningful properties. In this work, we review and investigate thirteen dimension reduction methods including linear and nonlinear, spectral, blind source separation, convex and non-convex methods and utilize the resulting embeddings to construct a mapping to quantities of interest via polynomial chaos expansions (PCE). We refer to the general proposed approach as manifold PCE (m-PCE), where manifold corresponds to the latent space resulting from any of the studied dimension reduction methods. To investigate the capabilities and limitations of these methods we conduct numerical tests for three physics-based systems (treated as black-boxes) having high-dimensional stochastic inputs of varying complexity modeled as both Gaussian and non-Gaussian random fields to investigate the effect of the intrinsic dimensionality of input data. We demonstrate both the advantages and limitations of the unsupervised learning methods and we conclude that a suitable m-PCE model provides a cost-effective approach compared to alternative algorithms proposed in the literature, including recently proposed expensive deep neural network-based surrogates and can be readily applied for high-dimensional UQ in stochastic PDEs.

arXiv:2202.02685

Blind source separation of baseband RF communication signals using mixed-signal matrix multiplication circuit

An 8 x 8 mixed-signal matrix multiplier architecture based on 64 hybrid capacitor-resistor multiplying digital to analogue converters implemented in a 65 nm CMOS technology was developed for the application of blind source separation of baseband RF signals. The integrated circuit has 13-bit resolution for each matrix weight and achieves a measured dynamic range of > 62 dB with a bandwidth of > 15 MHz and typical power dissipation of < 30 mW per matrix row. Separation of single-tone signal is measured to be better than 57 dBc.

arXiv:2201.06123공식 코드 · Papers with Code

Modeling the Repetition-based Recovering of Acoustic and Visual Sources with Dendritic Neurons

In natural auditory environments, acoustic signals originate from the temporal superimposition of different sound sources. The problem of inferring individual sources from ambiguous mixtures of sounds is known as blind source decomposition. Experiments on humans have demonstrated that the auditory system can identify sound sources as repeating patterns embedded in the acoustic input. Source repetition produces temporal regularities that can be detected and used for segregation. Specifically, listeners can identify sounds occurring more than once across different mixtures, but not sounds heard only in a single mixture. However, whether such a behaviour can be computationally modelled has not yet been explored. Here, we propose a biologically inspired computational model to perform blind source separation on sequences of mixtures of acoustic stimuli. Our method relies on a somatodendritic neuron model trained with a Hebbian-like learning rule which can detect spatio-temporal patterns recurring in synaptic inputs. We show that the segregation capabilities of our model are reminiscent of the features of human performance in a variety of experimental settings involving synthesized sounds with naturalistic properties. Furthermore, we extend the study to investigate the properties of segregation on task settings not yet explored with human subjects, namely natural sounds and images. Overall, our work suggests that somatodendritic neuron models offer a promising neuro-inspired learning strategy to account for the characteristics of the brain segregation capabilities as well as to make predictions on yet untested experimental settings.

arXiv:2201.05577

Unsupervised Sparse Unmixing of Atmospheric Trace Gases from Hyperspectral Satellite Data

In this letter, a new approach for the retrieval of the vertical column concentrations of trace gases from hyperspectral satellite observations, is proposed. The main idea is to perform a linear spectral unmixing by estimating the abundances of trace gases spectral signatures in each mixed pixel collected by an imaging spectrometer in the ultraviolet region. To this aim, the sparse nature of the measurements is brought to light and the compressive sensing paradigm is applied to estimate the concentrations of the gases' endemembers given by an a priori wide spectral library, including reference cross sections measured at different temperatures and pressures at the same time. The proposed approach has been experimentally assessed using both simulated and real hyperspectral dataset. Specifically, the experimental analysis relies on the retrieval of sulfur dioxide during volcanic emissions using data collected by the TROPOspheric Monitoring Instrument. To validate the procedure, we also compare the obtained results with the sulfur dioxide total column product based on the differential optical absorption spectroscopy technique and the retrieved concentrations estimated using the blind source separation.

arXiv:2112.13021

Noninvasive Fetal Electrocardiography: Models, Technologies and Algorithms

The fetal electrocardiogram (fECG) was first recorded from the maternal abdominal surface in the early 1900s. During the past fifty years, the most advanced electronics technologies and signal processing algorithms have been used to convert noninvasive fetal electrocardiography into a reliable technology for fetal cardiac monitoring. In this chapter, the major signal processing techniques, which have been developed for the modeling, extraction and analysis of the fECG from noninvasive maternal abdominal recordings are reviewed and compared with one another in detail. The major topics of the chapter include: 1) the electrophysiology of the fECG from the signal processing viewpoint, 2) the mathematical model of the maternal volume conduction media and the waveform models of the fECG acquired from body surface leads, 3) the signal acquisition requirements, 4) model-based techniques for fECG noise and interference cancellation, including adaptive filters and semi-blind source separation techniques, and 5) recent algorithmic advances for fetal motion tracking and online fECG extraction from few number of channels.

arXiv:2112.05694공식 코드 · Papers with Code

Unrolling PALM for sparse semi-blind source separation

Sparse Blind Source Separation (BSS) has become a well established tool for a wide range of applications - for instance, in astrophysics and remote sensing. Classical sparse BSS methods, such as the Proximal Alternating Linearized Minimization (PALM) algorithm, nevertheless often suffer from a difficult hyperparameter choice, which undermines their results. To bypass this pitfall, we propose in this work to build on the thriving field of algorithm unfolding/unrolling. Unrolling PALM enables to leverage the data-driven knowledge stemming from realistic simulations or ground-truth data by learning both PALM hyperparameters and variables. In contrast to most existing unrolled algorithms, which assume a fixed known dictionary during the training and testing phases, this article further emphasizes on the ability to deal with variable mixing matrices (a.k.a. dictionaries). The proposed Learned PALM (LPALM) algorithm thus enables to perform semi-blind source separation, which is key to increase the generalization of the learnt model in real-world applications. We illustrate the relevance of LPALM in astrophysical multispectral imaging: the algorithm not only needs up to $10^4-10^5$ times fewer iterations than PALM, but also improves the separation quality, while avoiding the cumbersome hyperparameter and initialization choice of PALM. We further show that LPALM outperforms other unrolled source separation methods in the semi-blind setting.

arXiv:2111.10574

Switching Independent Vector Analysis and Its Extension to Blind and Spatially Guided Convolutional Beamforming Algorithms

This paper develops a framework that can perform denoising, dereverberation, and source separation accurately by using a relatively small number of microphones. It has been empirically confirmed that Independent Vector Analysis (IVA) can blindly separate N sources from their sound mixture even with diffuse noise when a sufficiently large number (=M) of microphones are available (i.e., M>>N). However, the estimation accuracy seriously degrades as the number of microphones, or more specifically M-N (>=0), decreases. To overcome this limitation of IVA, we propose switching IVA (swIVA) in this paper. With swIVA, time frames of an observed signal with time-varying characteristics are clustered into several groups, each of which can be well handled by IVA using a small number of microphones, and thus accurate estimation can be achieved by applying IVA individually to each of the groups. Conventionally, a switching mechanism was introduced into a beamformer; however, no blind source separation algorithms with a switching mechanism have been successfully developed until this paper. In order to incorporate dereverberation capability, this paper further extends swIVA to blind Convolutional beamforming algorithm (swCIVA). It integrates swIVA and switching Weighted Prediction Error-based dereverberation (swWPE) in a jointly optimal way. We show that both swIVA and swCIVA can be optimized effectively based on blind signal processing, and that their performance can be further improved using a spatial guide for the initialization. Experiments show that both proposed methods largely outperform conventional IVA and its Convolutional beamforming extension (CIVA) in terms of objective signal quality and automatic speech recognition scores when using a relatively small number of microphones.

arXiv:2111.08858

A Normative and Biologically Plausible Algorithm for Independent Component Analysis

The brain effortlessly solves blind source separation (BSS) problems, but the algorithm it uses remains elusive. In signal processing, linear BSS problems are often solved by Independent Component Analysis (ICA). To serve as a model of a biological circuit, the ICA neural network (NN) must satisfy at least the following requirements: 1. The algorithm must operate in the online setting where data samples are streamed one at a time, and the NN computes the sources on the fly without storing any significant fraction of the data in memory. 2. The synaptic weight update is local, i.e., it depends only on the biophysical variables present in the vicinity of a synapse. Here, we propose a novel objective function for ICA from which we derive a biologically plausible NN, including both the neural architecture and the synaptic learning rules. Interestingly, our algorithm relies on modulating synaptic plasticity by the total activity of the output neurons. In the brain, this could be accomplished by neuromodulators, extracellular calcium, local field potential, or nitric oxide.

arXiv:2111.05120

A Deep Learning Technique using Low Sampling rate for residential Non Intrusive Load Monitoring

Individual device loads and energy consumption feedback is one of the important approaches for pursuing users to save energy in residences. This can help in identifying faulty devices and wasted energy by devices when left On unused. The main challenge is to identity and estimate the energy consumption of individual devices without intrusive sensors on each device. Non-intrusive load monitoring (NILM) or energy disaggregation, is a blind source separation problem which requires a system to estimate the electricity usage of individual appliances from the aggregated household energy consumption. In this paper, we propose a novel deep neural network-based approach for performing load disaggregation on low frequency power data obtained from residential households. We combine a series of one-dimensional Convolutional Neural Networks and Long Short Term Memory (1D CNN-LSTM) to extract features that can identify active appliances and retrieve their power consumption given the aggregated household power value. We used CNNs to extract features from main readings in a given time frame and then used those features to classify if a given appliance is active at that time period or not. Following that, the extracted features are used to model a generation problem using LSTM. We train the LSTM to generate the disaggregated energy consumption of a particular appliance. Our neural network is capable of generating detailed feedback of demand-side, providing vital insights to the end-user about their electricity consumption. The algorithm was designed for low power offline devices such as ESP32. Empirical calculations show that our model outperforms the state-of-the-art on the Reference Energy Disaggregation Dataset (REDD).

arXiv:2110.10546공식 코드 · Papers with Code

Trash or Treasure? An Interactive Dual-Stream Strategy for Single Image Reflection Separation

Single image reflection separation (SIRS), as a representative blind source separation task, aims to recover two layers, $\textit{i.e.}$, transmission and reflection, from one mixed observation, which is challenging due to the highly ill-posed nature. Existing deep learning based solutions typically restore the target layers individually, or with some concerns at the end of the output, barely taking into account the interaction across the two streams/branches. In order to utilize information more efficiently, this work presents a general yet simple interactive strategy, namely $\textit{your trash is my treasure}$ (YTMT), for constructing dual-stream decomposition networks. To be specific, we explicitly enforce the two streams to communicate with each other block-wisely. Inspired by the additive property between the two components, the interactive path can be easily built via transferring, instead of discarding, deactivated information by the ReLU rectifier from one stream to the other. Both ablation studies and experimental results on widely-used SIRS datasets are conducted to demonstrate the efficacy of YTMT, and reveal its superiority over other state-of-the-art alternatives. The implementation is quite simple and our code is publicly available at $\href{https://github.com/mingcv/YTMT-Strategy}{\textit{https://github.com/mingcv/YTMT-Strategy}}$.

arXiv:2110.06568

One to Multiple Mapping Dual Learning: Learning Multiple Sources from One Mixed Signal

Single channel blind source separation (SCBSS) refers to separate multiple sources from a mixed signal collected by a single sensor. The existing methods for SCBSS mainly focus on separating two sources and have weak generalization performance. To address these problems, an algorithm is proposed in this paper to separate multiple sources from a mixture by designing a parallel dual generative adversarial Network (PDualGAN) that can build the relationship between a mixture and the corresponding multiple sources to realize one-to-multiple cross-domain mapping. This algorithm can be applied to any mixed model such as linear instantaneous mixed model and convolutional mixed model. Besides, one-to-multiple datasets are created which including the mixtures and corresponding sources for this study. The experiment was carried out on four different datasets and tested with signals mixed in different proportions. Experimental results show that the proposed algorithm can achieve high performance in peak signal-to-noise ratio (PSNR) and correlation, which outperforms state-of-the-art algorithms.

arXiv:2109.09499

Machine learning methods for modelling and analysis of time series signals in geoinformatics

In this dissertation is provided a comparative analysis that evaluates the performance of several deep learning (DL) architectures on a large number of time series datasets of different nature and for different applications. Two main fruitful research fields are discussed here which were strategically chosen in order to address current cross disciplinary research priorities attracting the interest of geodetic community. The first problem is related to ionospheric Total Electron Content (TEC) modeling which is an important issue in many real time Global Navigation System Satellites (GNSS) applications. Reliable and fast knowledge about ionospheric variations becomes increasingly important. GNSS users of single frequency receivers and satellite navigation systems need accurate corrections to remove signal degradation effects caused by the ionosphere. Ionospheric modeling using signal processing techniques is the subject of discussion in the present contribution. The next problem under discussion is energy disaggregation which is an important issue for energy efficiency and energy consumption awareness. Reliable and fast knowledge about residential energy consumption at appliance level becomes increasingly important nowadays and it is an important mitigation measure to prevent energy wastage. Energy disaggregation or Nonintrusive load monitoring (NILM) is a single channel blind source separation problem where the task is to estimate the consumption of each electrical appliance given the total energy consumption. For both problems various deep learning models (DL) are proposed that cover various aspects of the problem under study, whereas experimental results indicate the proposed methods superiority compared to the current state of the art.

arXiv:2109.05425

HYPERION: Hyperspectral Penetrating-type Ellipsoidal Reconstruction for Terahertz Blind Source Separation

Terahertz (THz) technology has been a great candidate for applications, including pharmaceutic analysis, chemical identification, and remote sensing and imaging due to its non-invasive and non-destructive properties. Among those applications, penetrating-type hyperspectral THz signals, which provide crucial material information, normally involve a noisy, complex mixture system. Additionally, the measured THz signals could be ill-conditioned due to the overlap of the material absorption peak in the measured bands. To address those issues, we consider penetrating-type signal mixtures and aim to develop a blind hyperspectral unmixing (HU) method without requiring any information from a prebuilt database. The proposed HYperspectral Penetrating-type Ellipsoidal ReconstructION (HYPERION) algorithm is unsupervised, not relying on collecting extensive data or sophisticated model training. Instead, it is developed based on elegant ellipsoidal geometry under a very mild requirement on data purity, whose excellent efficacy is experimentally demonstrated.

arXiv:2108.09353

Temporally Nonstationary Component Analysis; Application to Noninvasive Fetal Electrocardiogram Extraction

Objective: Mixtures of temporally nonstationary signals are very common in biomedical applications. The nonstationarity of the source signals can be used as a discriminative property for signal separation. Herein, a semi-blind source separation algorithm is proposed for the extraction of temporally nonstationary components from linear multichannel mixtures of signals and noises. Methods: A hypothesis test is proposed for the detection and fusion of temporally nonstationary events, by using ad hoc indexes for monitoring the first and second order statistics of the innovation process. As proof of concept, the general framework is customized and tested over noninvasive fetal cardiac recordings acquired from the maternal abdomen, over publicly available datasets, using two types of nonstationarity detectors: 1) a local power variations detector, and 2) a model-deviations detector using the innovation process properties of an extended Kalman filter. Results: The performance of the proposed method is assessed in presence of white and colored noise, in different signal-to-noise ratios. Conclusion and Significance: The proposed scheme is general and it can be used for the extraction of nonstationary events and sample deviations from a presumed model in multivariate data, which is a recurrent problem in many machine learning applications.

arXiv:2107.14135

Modifications of FastICA in Convolutive Blind Source Separation

Convolutive blind source separation (BSS) is intended to recover the unknown components from their convolutive mixtures. Contrary to the contrast functions used in instantaneous cases, the spatial-temporal prewhitening stage and the para-unitary filters constraint are difficult to implement in a convolutive context. In this paper, we propose several modifications of FastICA to alleviate these difficulties. Our method performs the simple prewhitening step on convolutive mixtures prior to the separation and optimizes the contrast function under the diagonalization constraint implemented by single value decomposition (SVD). Numerical simulations are implemented to verify the performance of the proposed method.