Search

Abstract Background noises are usually treated as redundant or even harmful to voice conversion. Therefore, when converting noisy speech, a pretrained module of speech separation is usually deployed to estimate clean speech prior to the conversion. However, this can lead to speech distortion due to the mismatch between the separation module and the conversion one. In this paper, a noise-robust voice conversion model is proposed, where a user can choose to retain or to remove the background sounds freely. Firstly, a speech separation module with a dual-decoder structure is proposed, where two decoders decode the denoised speech and the background sounds, respectively. A bridge module is used to capture the interactions between the denoised speech and the background sounds in parallel layers through information exchanging. Subsequently, a voice conversion module with multiple encoders to convert the estimated clean speech from the speech separation model. Finally, the speech separation and voice conversion module are jointly trained using a loss function combining cycle loss and mutual information loss, aiming to improve the decoupling efficacy among speech contents, pitch, and speaker identity. Experimental results show that the proposed model obtains significant improvements in both subjective and objective evaluation metrics compared with the existing baselines. The speech naturalness and speaker similarity of the converted speech are 3.47 and 3.43, respectively.

Formononetin (FMNT) is a secondary metabolite of flavonoids abundant in legumes and graminaceous plants such as Astragalus mongholicus Bunge [Fabaceae; Astragali radix] and Avena sativa L. [Poaceae]. Astragalus is traditionally used in Asia countries such as China, Korea and Mongolia to treat inflammatory diseases, immune disorders and cancers. In recent years, inflammation and oxidative stress have been found to be associated with many diseases. A large number of pharmacological studies have shown that FMNT, an important bioactive metabolite of Astragalus, has a profoundly anti-inflammatory and antioxidant potential. This review focuses on providing comprehensive and up-to-date findings on the efficacy of the molecular targets and mechanisms involve of FMNT and its derivatives against inflammation and oxidative stress in both in vitro and in vivo. Relevant literature on FMNT against inflammation and oxidative stress between 2013 and 2023 were analyzed. FMNT has antioxidant and anti-inflammatory potential and shows mild or no toxicity in various diseases. Moreover, in the medical field, FMNT has shown potential in the prevention and treatment of cancers, neurological diseases, fibrotic diseases, allergic diseases, metabolic diseases, cardiovascular diseases, gastrointestinal diseases and autoimmune diseases. Thus, it is expected to be utilized in more products in the medical, food and cosmetic industries in the future.

Abstract Automatic speaker recognition is an important biometric authentication approach with emerging applications. However, recent research has shown its vulnerability on adversarial attacks. In this paper, we propose a new type of adversarial examples by generating imperceptible adversarial samples for targeted attacks on black-box systems of automatic speaker recognition. Waveform samples are created directly by solving an optimization problem with waveform inputs and outputs, which is more realistic in real-life scenario. Inspired by auditory masking, a regularization term adapting to the energy of speech waveform is proposed for generating imperceptible adversarial perturbations. The optimization problems are subsequently solved by differential evolution algorithm in a black-box manner which does not require any knowledge on the inner configuration of the recognition systems. Experiments conducted on commonly used data sets, LibriSpeech and VoxCeleb, show that the proposed methods have successfully performed targeted attacks on state-of-the-art speaker recognition systems while being imperceptible to human listeners. Given the high SNR and PESQ scores of the yielded adversarial samples, the proposed methods deteriorate less on the quality of the original signals than several recently proposed methods, which justifies the imperceptibility of adversarial samples.

Abstract In skeleton‐based action recognition, the graph convolutional network (GCN) has achieved great success. Modelling skeleton data in a suitable spatial‐temporal way and designing the adjacency matrix are crucial aspects for GCN‐based methods to capture joint relationships. In this study, we propose the spatial‐temporal slowfast graph convolutional network (STSF‐GCN) and design the adjacency matrices for the skeleton data graphs in STSF‐GCN. STSF‐GCN contains two pathways: (1) the fast pathway is in a high frame rate, and joints of adjacent frames are unified to build ‘small’ spatial‐temporal graphs. A new spatial‐temporal adjacency matrix is proposed for these ‘small’ spatial‐temporal graphs. Ablation studies verify the effectiveness of the proposed adjacency matrix. (2) The slow pathway is in a low frame rate, and joints from all frames are unified to build one ‘big’ spatial‐temporal graph. The adjacency matrix for the ‘big’ spatial‐temporal graph is obtained by computing self‐attention coefficients of each joint. Finally, outputs from two pathways are fused to predict the action category. STSF‐GCN can efficiently capture both long‐range and short‐range spatial‐temporal joint relationships. On three datasets for skeleton‐based action recognition, STSF‐GCN can achieve state‐of‐the‐art performance with much less computational cost.

Automatic speaker verification (ASV) is an emerging biometric verification technique with more and more applications. However, both verification accuracy and anti-spoofing should be considered carefully before putting ASV into practice, where anti-spoofing is also called replay detection in which voice is recorded, stored and replayed to deceive ASV systems. Cascaded decision of anti-spoofing and ASV is a straightforward solution to tackle the two issues. In this paper, joint decision of anti-spoofing and ASV was investigated in a multi-task learning framework with contrastive loss in order to improve the cascaded decision approach. A modified triplet loss was firstly constructed to supervise deep neural networks to extract embedding vectors containing information of both speaker identity and spoofing. The embedding vectors were subsequently taken as input features by back-end classifiers towards speaker and spoofing classification. The experimental results on both ASVspoof 2017 and ASVspoof 2019 showed that the proposed joint decision approach with triplet loss outperformed the corresponding baselines, a recent work on joint decision with Gaussian back-end fusion and our previous joint decision approach with cross-entropy loss.

Abstract Background Common carotid artery (CCA) occlusion (CCAO) is a rare condition. Owing to collateral circulation, ipsilateral internal carotid artery (ICA) and external carotid artery (ECA) are often patent. Methods This study included 16 patients with unilateral CCAO and patent ipsilateral ICA and ECA. The pathways which supplied ICA were investigated by digital subtraction angiography (DSA), transcranial Doppler (TCD), magnetic resonance angiography (MRA) and computed tomography angiography (CTA). Results In all 16 patients, TCD found antegrade blood flow in ipsilateral ICA, which was supplied by retrograde blood flow in ipsilateral ECA through carotid bifurcation. We call this phenomenon “ICA steal”. DSA and CTA discovered four pathways of ICA steal, including 1) ipsilateral vertebral artery – occipital artery – ECA – ICA, 2) ipsilateral thyrocervical trunk or costocervical trunk – ascending cervical artery or deep cervical artery – occipital artery – ECA – ICA, 3) contralateral ECA – contralateral superior thyroid artery – ipsilateral superior thyroid artery – ipsilateral ECA – ICA, and 4) ipsilateral thyrocervical trunk – inferior thyroid artery – superior thyroid artery – ECA – ICA. Conclusions ICA is possible to be patent and supplied by several collateral circulation pathways after CCAO.

In recent years, neural network-based voice conversion methods have been rapidly developed, and many different models and neural networks have been applied in parallel voice conversion. However, the over-smoothing of parametric methods [e.g., bidirectional long short-term memory (BLSTM)] and the waveform collapse of neural vocoders (e.g., WaveNet) still have negative impacts on the quality of the converted voices. To overcome this problem, we propose a BLSTM and WaveNet-based voice conversion method cooperated with waveform collapse suppression by post-processing. This method firstly uses BLSTM to convert the acoustic features between parallel speakers, and then synthesizes pre-converted voice with WaveNet. Subsequently, several alternative iterations of BLSTM post-processing is performed, and the final converted voice is generated by WaveNet. The proposed method can directly generate converted audio waveforms and avoid the waveform-collapsed speech caused by a single WaveNet generation effectively. The experimental results indicate that acoustic features trained by using the BLSTM network could achieve better results than conventional baselines. From our experiments on VCC2018, the usage of WaveNet could alleviate the problem of over-smoothing, which contributes to improving the similarity and naturalness of the final results of voice conversion.

Desirable properties of extensions of non-negative matrix factorization (NMF) include robustness in the presence of noises and outliers, ease of implementation, the guarantee of convergence, operation in an automatic fashion that trades off the balance between data approximation and model simplicity well, and the capability to model the inherently sequential structure of time-series signals. The state-of-the-art methods typically have only a subset of these aforementioned properties and seldom simultaneously possess them all. In this paper, we propose a novel approach that provides all these desirable properties by extending the automatic relevance determination framework in NMF from Tan and Févotte. Starting from an objective function derived from the maximum a posterior estimation of a Bayesian model, we develop majorization-minimization algorithms that work effectively to determine the correct model order, regardless of the impact of noise and outliers. Subsequently, we give a rigorous convergence analysis of the proposed algorithms. Moreover, convolutive bases are also incorporated in the basic model so that it is able to capture the richness of temporal continuity. We perform experiments on both synthetic and real-world data sets to show the efficiency and robustness of our approach.

Based on the concepts of pseudocomplement of L-subsets and the implication operator where L is a completely distributive lattice with order-reversing involution, the definition of countable RL-fuzzy compactness degree and the Lindelöf property degree of an L-subset in RL-fuzzy topology are introduced and characterized. Since L-fuzzy topology in the sense of Kubiak and Šostak is a special case of RL-fuzzy topology, the degrees of RL-fuzzy compactness and the Lindelöf property are generalizations of the corresponding degrees in L-fuzzy topology.

This paper aims to make a combination between the quantum B-algebras (briefly, X-As) and two interesting theories (e.g., soft set theory and fuzzy soft set theory). Firstly, we propose the novel notions of soft quantum B-algebras (briefly, SℚB-As), a soft deductive system of ℚB-As, and deducible soft quantum B-algebras (briefly, DSℚB-As). Then, we discuss the relationship between SℚB-As and DSℚB-As. Furthermore, we investigate the union and intersection operations of DSℚB-As. Secondly, we introduce the notions of a fuzzy soft quantum B-algebras (briefly, FSℚB-As), a fuzzy soft deductive system of ℚB-As, and present some characterizations of FSℚB-As, along with several examples. Finally, we explain the basic properties of homomorphism image of FSℚB-As.

Twitter is a virtual social network where people share their posts and opinions about the current situation, such as the coronavirus pandemic. It is considered the most significant streaming data source for machine learning research in terms of analysis, prediction, knowledge extraction, and opinions. Sentiment analysis is a text analysis method that has gained further significance due to social networks’ emergence. Therefore, this paper introduces a real-time system for sentiment prediction on Twitter streaming data for tweets about the coronavirus pandemic. The proposed system aims to find the optimal machine learning model that obtains the best performance for coronavirus sentiment analysis prediction and then uses it in real-time. The proposed system has been developed into two components: developing an offline sentiment analysis and modeling an online prediction pipeline. The system has two components: the offline and the online components. For the offline component of the system, the historical tweets’ dataset was collected in duration 23/01/2020 and 01/06/2020 and filtered by #COVID-19 and #Coronavirus hashtags. Two feature extraction methods of textual data analysis were used, n-gram and TF-ID, to extract the dataset’s essential features, collected using coronavirus hashtags. Then, five regular machine learning algorithms were performed and compared: decision tree, logistic regression, k-nearest neighbors, random forest, and support vector machine to select the best model for the online prediction component. The online prediction pipeline was developed using Twitter Streaming API, Apache Kafka, and Apache Spark. The experimental results indicate that the RF model using the unigram feature extraction method has achieved the best performance, and it is used for sentiment prediction on Twitter streaming data for coronavirus.

Bone-conducted (BC) speech can be used to communicate in a very high noise environment. In this paper, a method of improving the quality of BC speech is presented. The speech signal of a speaker is passed through a novel dictionary representation-based encoder-decoder model. In the encoder, our designed non-negative and sparse long short-term memory (LSTM) recurrent neural network is deployed to generate combination coefficients on the dictionary established by sparse non-negative matrix factorization. Then, the decoder is designed and utilized to enhance the dictionary representation based on local attention mechanism. Two optimizers are adopted when training the model as a whole and the encoder is pre-trained individually to make the convergence faster. In experiments, we compare the proposed method with direct transformations via DNN and LSTM networks, and numerous criteria are used for evaluation. Objective and subjective results demonstrate that our method behaves better and achieves satisfactory performance even when coping with some challenging cases.

Non-acoustic sensors are widely used in speech signal processing tasks, and their immunity to the background acoustic noise shows great benefits to traditional speech enhancement. To avoid using acoustic speech disturbed by strong noise, spectra mapping from throat microphone (TM) speech to acoustic microphone (AM) speech has been studied. However, there is a distinguished difference between the spectra of the two kinds of speech, and the mapping relationship is different in the low-band and high-band spectra, which limits the performance of the traditional full-band spectra mapping model. In this paper, to improve the perceived quality and intelligibility of TM speech, we investigate the low-band and high-band spectral structure between TM and AM speech, respectively, and propose a spectra-based band-division mapping framework for TM speech enhancement based on the investigation. In the framework, the low-band target spectra of AM speech are mapped based on the equalization method, and the high-band spectra are mapped from the full-band TM speech spectra, which are lack of high-frequency components. The overall framework can be viewed as a combination of spectra equalization in the low band and spectra generation in the high band. Both the objective and subjective evaluation results show clear advantages over the existing TM speech enhancement method based on the full-band spectra mapping.

Regarding the sequence learning of neural networks, there exists a problem of how to capture long-term dependencies and alleviate the gradient vanishing phenomenon. To manage this problem, we proposed a neural network with random connections via a scheme of a neural architecture search. First, a dense network was designed and trained to construct a search space, and then another network was generated by random sampling in the space, whose skip connections could transmit information directly over multiple periods and capture long-term dependencies more efficiently. Moreover, we devised a novel cell structure that required less memory and computational power than the structures of long short-term memories (LSTMs), and finally, we performed a special initialization scheme on the cell parameters, which could permit unhindered gradient propagation on the time axis at the beginning of training. In the experiments, we evaluated four sequential tasks: adding, copying, frequency discrimination, and image classification; we also adopted several state-of-the-art methods for comparison. The experimental results demonstrated that our proposed model achieved the best performance.

This paper presents a noise-robust voice conversion method with high-quefrency boosting via sub-band cepstrum conversion and fusion based on the bidirectional long short-term memory (BLSTM) neural networks that can convert parameters of vocal tracks of a source speaker into those of a target speaker. With the implementation of state-of-the-art machine learning methods, voice conversion has achieved good performance given abundant clean training data. However, the quality and similarity of the converted voice are significantly degraded compared to that of a natural target voice due to various factors, such as limited training data and noisy input speech from the source speaker. To address the problem of noisy input speech, an architecture of voice conversion with statistical filtering and sub-band cepstrum conversion and fusion is introduced. The impact of noises on the converted voice is reduced by the accurate reconstruction of the sub-band cepstrum and the subsequent statistical filtering. By normalizing the mean and variance of the converted cepstrum to those of the target cepstrum in the training phase, a cepstrum filter was constructed to further improve the quality of the converted voice. The experimental results showed that the proposed method significantly improved the naturalness and similarity of the converted voice compared to the baselines, even with the noisy inputs of source speakers.

Even though audio replay detection has improved in recent years, its performance is known to severely deteriorate with the existence of strong background noises. Given the fact that different frames of an utterance have different impacts on the performance of spoofing detection, this paper introduces attention-based long short-term memory (LSTM) to extract representative frames for spoofing detection in noisy environments. With this attention mechanism, the specific and representative frame-level features will be automatically selected by adjusting their weights in the framework of attention-based LSTM. The experiments, conducted using the ASVspoof 2017 dataset version 2.0, show that the equal error rate (EER) of the proposed approach was about 13% lower than the constant Q cepstral coefficients-Gaussian mixture model (CQCC-GMM) baseline in noisy environments with four different signal-to-noise ratios (SNR). Meanwhile, the proposed algorithm also improved the performance of traditional LSTM on audio replay detection systems in noisy environments. Experiments using bagging with different frame lengths were also conducted to further improve the proposed approach.