Speech spoofing countermeasures based on source voice analysis and machine learning techniques

This paper presents a robust user authentication system by gleaning raw mouse movement data. The data was collected using a publicly available tool called Recording User Input (RUI) from 23 subjects analyzed for three types of mouse actions - Mouse Move, Point-and-Click on Left or Right mouse button, and Drag-and-Drop. Samples are broken down to unit blocks comprising a certain number of actions and from each block seventy-four features are extracted to construct feature vectors. The proposed system was rigorously tested against public benchmark data. Experiment results generated by using the Support Vector Machine (SVM) classifier shows a False Rejection Rate (FRR) of 1.1594 % and a False Acceptance Rate (FAR) of 1.9053 % when the block size was set for 600 actions. After reducing dimensions using Principle Component Analysis (PCA), SVM classifier shows FRR of 1.2081 % and FAR of 2.3604 %. Compared with the existing methods based on mouse movements, our method shows significantly lower error rates, which we opine are viable enough to become an alternate to conventional authentication systems.

 Forecasting stream flow accurately is essential for managing water resources, preventing flooding, and designing the environment for intricate watersheds. This study conducts a comprehensive assessment of streamflow simulation models—Multilayer Perceptron (MLP), Extreme Learning Machine (ELM), Support Vector Machine (SVM), and the Soil and Water Assessment Tool (SWAT) hydrological model—covering the period from 1985 to 2018 in the Astore Basin, Pakistan. The GMRC-WAPDA and SWHP-WAPDA provided daily streamflow data from the Doyian gauging station in the Astor River Basin, as well as meteorological data collected from two automatic weather stations (AWS) located at Rama, Rattu and Astor. A total number of four soil classes (lithosols, calcaric, gleysols and fluvisol) was observed in the basin. The primary objective was to comprehensively assess the predictive performances of these models across distinct time segments and gauge their reliability in simulating streamflow dynamics. The study commenced with examining the SWAT model's performance, utilizing the NSE, PBIAS, R2, and RMSE metrics during calibration (1985–2000) and validation (2001–2009) periods. While the SWAT model effectively estimated streamflow, it exhibited limitations in accurately predicting peak and low-flow conditions. Subsequently, the machine learning models (MLP, ELM, and SVM) were scrutinized concerning their performance metrics—R2, NSE, PBIAS, and RMSE—across training (1985–1995), validation (1996–2005), and testing (2005–2009) datasets. ELM displayed superior performance during the training phase, boasting a remarkable R2 of 0.94, followed by SVM and MLP. MLP showcased consistent strength in validation, maintaining an R2 of 0.73, while SVM followed with an R2 of 0.71. Despite their merits, none of the models precisely replicated observed streamflow patterns, as evidenced by the discrepancies between the observed flow and the SWAT model's simulations. This emphasizes the necessity for ongoing refinement and validation to enhance predictive accuracy and ensure closer alignment with real-world hydrological dynamics. This extensive comparative analysis offers critical insights into the nuances of MLP, ELM, SVM, and SWAT model performances, highlighting their varied strengths and limitations across distinct temporal segments. It underscores the importance of continual refinement and validation to improve predictive capabilities, which is essential for accurate streamflow simulations and effective water resource management in the Astor Basin and similar hydrological contexts.

With the growing number of devices that use voice biometric system for verification purposes, it becomes mandatory to accurately verify the target speaker. Moreover, intruders use different spoofing attacks to fool an automatic speaker verification (ASV) system. Voice replay attack being the easiest to generate among other spoofing attacks is frequently employed in front of the ASV systems of these devices to get access to the system for malicious purposes. Thus, we need to develop robust systems that are capable of classifying the target speaker and differentiate between the bonafide and replay voice. To combat the vulnerabilities of ASV systems, we proposed an automatic speaker verification and replay attack detection (ASVRAD) system. For this purpose, we proposed a novel audio feature i.e., Glottal Flow Gammatone Cepstrum Coefficients (GLGTCC), an enhanced version of Glottal Flow Cepstrum Coefficients (GLFCC) for accurate classification of the bonafide speakers and voice replay attacks detection. Our GLGTCC features can reliably capture the speaker-specific attributes from their vocal excitation patterns in the audio. Later, we employed a multiclass support vector machine (SVM) to classify the bonafide speakers, and a binary SVM for classification of bonafide and replay audios. Experimental results of our system on the standard ASVspoof 2019 PA dataset achieved the lowest EER of 3.00% for speaker verification and 18.38% on voice replay attacks over the comparative methods.

Bond rating is regarded as an important event for measuring financial risk of companies and for determining the investment returns of investors. As a result, it has been a popular research topic for researchers to predict companies' credit ratings by applying statistical and machine learning techniques. The statistical techniques, including multiple regression, multiple discriminant analysis (MDA), logistic models (LOGIT), and probit analysis, have been traditionally used in bond rating. However, one major drawback is that it should be based on strict assumptions. Such strict assumptions include linearity, normality, independence among predictor variables and pre-existing functional forms relating the criterion variablesand the predictor variables. Those strict assumptions of traditional statistics have limited their application to the real world. Machine learning techniques also used in bond rating prediction models include decision trees (DT), neural networks (NN), and Support Vector Machine (SVM). Especially, SVM is recognized as a new and promising classification and regression analysis method. SVM learns a separating hyperplane that can maximize the margin between two categories. SVM is simple enough to be analyzed mathematical, and leads to high performance in practical applications. SVM implements the structuralrisk minimization principle and searches to minimize an upper bound of the generalization error. In addition, the solution of SVM may be a global optimum and thus, overfitting is unlikely to occur with SVM. In addition, SVM does not require too many data sample for training since it builds prediction models by only using some representative sample near the boundaries called support vectors. A number of experimental researches have indicated that SVM has been successfully applied in a variety of pattern recognition fields. However, there are three major drawbacks that can be potential causes for degrading SVM's performance. First, SVM is originally proposed for solving binary-class classification problems. Methods for combining SVMs for multi-class classification such as One-Against-One, One-Against-All have been proposed, but they do not improve the performance in multi-class classification problem as much as SVM for binary-class classification. Second, approximation algorithms (e.g. decomposition methods, sequential minimal optimization algorithm) could be used for effective multi-class computation to reduce computation time, but it could deteriorate classification performance. Third, the difficulty in multi-class prediction problems is in data imbalance problem that can occur when the number of instances in one class greatly outnumbers the number of instances in the other class. Such data sets often cause a default classifier to be built due to skewed boundary and thus the reduction in the classification accuracy of such a classifier. SVM ensemble learning is one of machine learning methods to cope with the above drawbacks. Ensemble learning is a method for improving the performance of classification and prediction algorithms. AdaBoost is one of the widely used ensemble learning techniques. It constructs a composite classifier by sequentially training classifiers while increasing weight on the misclassified observations through iterations. The observations that are incorrectly predicted by previous classifiers are chosen more often than examples that are correctly predicted. Thus Boosting attempts to produce new classifiers that are better able to predict examples for which the current ensemble's performance is poor. In this way, it can reinforce the training of the misclassified observations of the minority class. This paper proposes a multiclass Geometric Mean-based Boosting (MGM-Boost) to resolve multiclass prediction problem. Since MGM-Boost introduces the notion of geometric mean into AdaBoost, it can perform learning process considering the geometric mean-based accuracy and errors of multiclass. This study applies MGM-Boost to the real-world bond rating case for Korean companies to examine the feasibility of MGM-Boost. 10-fold cross validations for threetimes with different random seeds are performed in order to ensure that the comparison among three different classifiers does not happen by chance. For each of 10-fold cross validation, the entire data set is first partitioned into tenequal-sized sets, and then each set is in turn used as the test set while the classifier trains on the other nine sets. That is, cross-validated folds have been tested independently of each algorithm. Through these steps, we have obtained the results for classifiers on each of the 30 experiments. In the comparison of arithmetic mean-based prediction accuracy between individual classifiers, MGM-Boost (52.95%) shows higher prediction accuracy than both AdaBoost (51.69%) and SVM (49.47%). MGM-Boost (28.12%) also shows the higher prediction accuracy than AdaBoost (24.65%) and SVM (15.42%)in terms of geometric mean-based prediction accuracy. T-test is used to examine whether the performance of each classifiers for 30 folds is significantly different. The results indicate that performance of MGM-Boost is significantly different from AdaBoost and SVM classifiers at 1% level. These results mean that MGM-Boost can provide robust and stable solutions to multi-classproblems such as bond rating.

With the development of speech synthesis technology, automatic speaker verification (ASV) systems have encountered the serious challenge of spoofing attacks. In order to improve the security of ASV systems, many antispoofing countermeasures have been developed. In the front-end domain, much research has been conducted on finding effective features which can distinguish spoofed speech from genuine speech and the published results show that dynamic acoustic features work more effectively than static ones. In the back-end domain, Gaussian mixture model (GMM) and deep neural networks (DNNs) are the two most popular types of classifiers used for spoofing detection. The log-likelihood ratios (LLRs) generated by the difference of human and spoofing log-likelihoods are used as spoofing detection scores. In this paper, we train a five-layer DNN spoofing detection classifier using dynamic acoustic features and propose a novel, simple scoring method only using human log-likelihoods (HLLs) for spoofing detection. We mathematically prove that the new HLL scoring method is more suitable for the spoofing detection task than the classical LLR scoring method, especially when the spoofing speech is very similar to the human speech. We extensively investigate the performance of five different dynamic filter bank-based cepstral features and constant Q cepstral coefficients (CQCC) in conjunction with the DNN-HLL method. The experimental results show that, compared to the GMM-LLR method, the DNN-HLL method is able to significantly improve the spoofing detection accuracy. Compared with the CQCC-based GMM-LLR baseline, the proposed DNN-HLL model reduces the average equal error rate of all attack types to 0.045%, thus exceeding the performance of previously published approaches for the ASVspoof 2015 Challenge task. Fusing the CQCC-based DNN-HLL spoofing detection system with ASV systems, the false acceptance rate on spoofing attacks can be reduced significantly.

Security assessment and classification are the major concerns in real-time operation of electric power systems. This paper proposes a multiclass support vector machine (SVM) classifier for static and transient security assessment and classification. A straightforward and quick procedure called the sequential forward selection method is used for a feature selection process. The security status of any given operating condition is classified into four modes, viz., secure, critically secure, insecure, and highly insecure, based on the computation of a security index. The proposed SVM-based pattern classifier system is implemented and tested on standard benchmark systems. The simulation results of the multiclass SVM classifier are compared with least-squares, probabilistic neural network, extreme learning machine, and extreme SVM classifiers. The feasibility of implementation of the proposed classifier system for online security evaluation is also discussed.

The study was intended to eliminate the noise in three-dimensional transvaginal ultrasound (3D-TVS) images and improve the diagnostic accuracy in intrauterine adhesion (IUA). The extreme learning machine (ELM) algorithm was introduced first for statement. One hundred and thirty cases of suspected IUA patients were taken as the research subjects. The denoising effects of ELM algorithm were evaluated in terms of mean square errors (MSE), peak signal-to-noise ratio (PSNR), and running time, and its diagnostic efficiency of IUA was identified from precise, specificity, and sensitivity. Furthermore, the support vector machine (SVM) algorithm was introduced for comparison. It was found that the MSE and PSNR of the ELM algorithm were 0.0021 and 64.5, respectively, and its average operation time was 11.22 ± 0.89s, that the MSE values of SVM algorithm and ELM algorithm were 0.0045 and 0.0021 and the PSNR values were 52.3 and 64.5, respectively, and that the average running time of SVM algorithm was 16.35 ± 1.33s, and the average running time of ELM algorithm was 11.22 ± 0.89s, superior to the SVM algorithm in denoising effects. Moreover, the ELM algorithm showed excellent diagnostic efficiency for patients with various degrees of IUA. In conclusion, ELM can effectively eliminate noise in 3D-TVS images and demonstrates excellent diagnostic efficiency on IUA, which is worthy of clinical application.

Playback attack is an approach for acquiring the unauthorized access to the automatic speaker verification (SV) system by using targets' pre-recorded speech samples. It is now a widely acknowledged fact that state-of-the-art SV systems are quite vulnerable to playback attacks, and developing countermeasures are in progress. In that direction most of the attempts try to capture the trace of the playback device characteristics by using spectral related information, that remain almost intact in playback signals, resulting less improvements in false acceptance rate. In this work we explore the use of linear prediction (LP) residual signal to counter the playback attacks. We observed that, as compared to spectral related vocal-tract features, the parametric representation of the LP residual signal shows relatively more robustness against playback attacks. In future, we aim to model the LP residual signal explicitly and develop countermeasures to playback attacks.

This paper proposes a face recognition algorithm based on the combination of local binary pattern (LBP) texture features and extreme learning machine (ELM). The face image is divided into several regions, and the LBP features are extracted from these regions and combined together to form a feature vector which will be the input data of ELM. It shows that ELM performs well in classification applications, and ELM and support vector machine (SVM) are equivalent from the optimization point of view. But ELM has milder optimization constraints and much less training time. Our experiments are carried out on two well-known face databases, and the results show that compared with compared to PCA+NN, PCA+SVM and PCA+ELM the proposed method can achieve higher recognition rates.

Robustness of Extreme Learning Machine in the prediction of hydrological flow series

Automatic Speaker Verification (ASV) systems are widely used for voice-based identity verification, but recent studies have shown that they are vulnerable to both over-the-line (OTL) and over-the-air (OTA) adversarial attacks. Although various detection methods have been proposed, their robustness against diverse OTA attack scenarios has not been systematically assessed. In this work, we present AdvSV 2.0, a large-scale, open-source dataset for adversarial attacks on ASV systems. AdvSV 2.0 contains 628k utterances (800 hours) spanning 8 ASV models, 4 attack methods, and 9 physical playback-recording device combinations, and includes both OTL and OTA adversarial examples. To better connect digital adversarial optimization with the physical OTA channel, we further propose an OTA attack generation method based on a Neural Replay Simulator (NRS), which simulates replay-channel distortions and improves OTA attack success rates by up to 33.4% after air transmission. To defend against these attacks, we propose CODA-OCC, a COntrastive Domain-Aligned One-Class Classifier. Built on a one-class classification paradigm, CODA-OCC is trained only on bona fide speech and does not rely on any white-box assumption about the ASV backend. Experimental results on AdvSV 2.0 show that CODA-OCC achieves an EER of 11.2% and an AUC of 0.95, outperforming the purification-based baseline by 8.6% absolute EER, and providing a robust foundation for studying OTA adversarial robustness in ASV.

With the development of spoofing technologies, automatic speaker verification (ASV) systems have encountered serious challenges on security. In order to address this problem, many anti-spoofing countermeasures have been explored. There are two intuitive recipes to protect an ASV system from spoofing. The first one is to use a cascaded structure where spoofing detection is performed firstly and ASV is subsequently conducted only on the attempts which have passed the spoofing detection. The other one is to perform spoofing detection and ASV jointly. The discriminate reliably of the joint system has been proven to be more advantageous than cascaded systems with traditional methods, not only in accuracy, but also in convenience and computational efficiency. In this paper, we proposed a multi-task learning approach based on deep neural network to make a joint system of ASV and anti-spoofing. The performance of different acoustic features and structures of deep neural networks has been investigated on the ASVspoof 2017 version 2.0 dataset. The experimental results showed that the joint equal error rate (EER) of our approach was reduced by 0.55% compared to a joint system with Gaussian back-end fusion baseline.

Microcalcification (MC) clusters in mammograms can be important early signs of breast cancer in women. Accurate detection of MC clusters is an important but challenging problem. In this paper, we propose the use of a recently developed machine learning technique -- relevance vector machine (RVM) -- for automatic detection of MCs in digitized mammograms. RVM is based on Bayesian estimation theory, and as a feature it can yield a decision function that depends on only a very small number of so-called <i>relevance vectors</i>. We formulate MC detection as a supervised-learning problem, and use RVM to classify if an MC object is present or not at each location in a mammogram image. MC clusters are then identified by grouping the detected MC objects. The proposed method is tested using a database of 141 clinical mammograms, and compared with a support vector machine (SVM) classifier which we developed previously. The detection performance is evaluated using the free-response receiver operating characteristic (FROC) curves. It is demonstrated that the RVM classifier matches closely with the SVM classifier in detection performance, and does so with a much sparser kernel representation than the SVM classifier. Consequently, the RVM classifier greatly reduces the computational complexity, making it more suitable for real-time processing of MC clusters in mammograms.

The problem with development of the support vector machine (SVM) classifiers using modified particle swarm optimization (PSO) algorithm and their ensembles has been considered. Solving this problem would allow fulfilling the high-precision data classification, especially Big Data classification, with the acceptable time expenditures. The modified PSO algorithm conducts a simultaneous search of the type of kernel functions, the parameters of the kernel function and the value of the regularization parameter for the SVM classifier. The idea of particles' «regeneration» served as the basis for the modified PSO algorithm. In the implementation of this algorithm, some particles change the type of their kernel function to the one which corresponds to the particle with the best value of the classification accuracy. The offered PSO algorithm allows reducing the time expenditures for the developed SVM classifiers, which is very important for Big Data classification problem. In most cases such SVM classifier provides the high quality of data classification. In exceptional cases the SVM ensembles based on the decorrelation maximization algorithm for the different strategies of the decision-making on the data classification and the majority vote rule can be used. Also, the two-level SVM classifier has been offered. This classifier works as the group of the SVM classifiers at the first level and as the SVM classifier on the base of the modified PSO algorithm at the second level. The results of experimental studies confirm the efficiency of the offered approaches for Big Data classification.

Various myoelectric prostheses controlled by electromyography (EMG) signals have been developed. However, there have been few studies that provide fast and accurate methods to predict handgrip force from EMG signals. Rapid and precise handgrip force prediction is required, especially for the real-time control system of myoelectric prostheses. In this study, extreme learning machine (ELM) is applied to predict handgrip force from surface EMG signals of forearm muscles. Furthermore, ELM is compared with support vector machine (SVM) and multiple nonlinear regression (MNLR). The below 10 % of the surface EMG and handgrip force signals were cut away, and then the root mean square feature extracted from the modified surface EMG signals was taken as input vector for these three kinds of predicting mechanisms. For the testing dataset, ELM achieved a slightly larger root mean squared error than SVM did and a smaller one than MNLR did. Meanwhile, all three methods showed high correlation coefficients. For the total processing time, ELM and MNLR consumed much less time than SVM did. Experimental results demonstrate that ELM possesses a relatively good accuracy and little consumed time, although SVM is effective for handgrip force estimation in terms of accuracy. Overall, ELM has a promising potential for predicting handgrip force rapidly and precisely.