Problem Of Speech Recognition Research Articles

Improvement and Application of Hale’s Dynamic Time Warping Algorithm

Due to the different generation and propagation mechanisms of P- and S-waves, there may be significant differences in the seismic data collected by the two, which poses a great obstacle to the time domain matching of P- and S-waves in multiwave exploration. Furthermore, the quality and accuracy of the matching effect will directly affect the subsequent multiwave joint inversion and interpretation effect. Therefore, the study of P and S-wave-matching methods plays a crucial role in seismic exploration. In 2013, Hale improved the classical Dynamic Time Warping (DTW) algorithm applied to solve the problem of speech recognition, and obtained the DTW algorithm suitable for solving the matching of P-waves and S-waves. The seismic wave-matching results generated by this algorithm are horizontal discontinuous (different trajectories) and need further processing. This study analyses the algorithm based on simulations of seismic waves using Ricker wavelets. In response to existing problems, this paper proposes strategies to improve the DTW algorithm. The algorithm in this study significantly improved the continuity of the registration results of the actual seismic wave data in the horizontal direction (different traces).

Integrated End-to-End automatic speech recognition for languages for agglutinative languages

The relevance of the problem of automatic speech recognition lies in the lack of research for low-resource languages, stemming from limited training data and the necessity for new technologies to enhance efficiency and performance. The purpose of this work was to study the main aspects of integrated end-to-end speech recognition and the use of modern technologies in the natural processing of agglutinative languages, including Kazakh. In this article, the study of language models was carried out using comparative, graphic, statistical and analytical-synthetic methods, which were used in combination. This paper addresses automatic speech recognition (ASR) in agglutinative languages, particularly Kazakh, through a unified neural network model that integrates both acoustic and language modeling. Employing advanced techniques like connectionist temporal classification and attention mechanisms, the study focuses on effective speech-to-text transcription for languages with complex morphologies. Transfer learning from high-resource languages helps mitigate data scarcity in languages such as Kazakh, Kyrgyz, Uzbek, Turkish, and Azerbaijani. The research assesses model performance, underscores ASR challenges, and proposes advancements for these languages. It includes a comparative analysis of phonetic and word-formation features in agglutinative Turkic languages, using statistical data. The findings aid further research in linguistics and technology for enhancing speech recognition and synthesis, contributing to voice identification and automation processes.

Improved Speaker Recognition System Using Automatic Lip Recognition

The paper is focused on the relevant problem of speech recognition using additional sources besides the voice itself, in conditions in which the quality or availability of audio information is inadequate (for example, in the presence of noise or additional speakers). This is achieved by using automatic lip recognition (ARL) methods, which rely on non-acoustic biosignals generated by the human body during speech production. Among the applications of this approach are medical applications, as well as processing voice commands in languages with poor audio conditions. The aim of this work is to create a system for speech recognition based on a combination of speaker lip recognition (SSI) and context prediction. To achieve this goal, the following tasks were performed: to substantiate the systems for recognizing voice commands of a silent voice interface (SSI) based on a combination of two neural network architectures, to implement a model for recognizing visemes based on the CNN neural network architecture and an encoder-decoder architecture for the LSTM neural recurrent network model for analyzing and predicting the context of a speaker’s speech. The developed system was tested on a chosen dataset. The results show that the recognition error in different conditions averages from 4,34% to 5,12% for CER and from 5,52% to 6,06% for WER for the proposed ALR system in 7 experiments, which is an advantage over the LipNet project, which additionally processes audio data for the original without noise.

Design of Internet Platform for English Teaching in Colleges and Universities under the Background of Big Data

Abstract This paper constructs an English teaching platform based on a speech recognition algorithm and Internet technology, adopts a three-tier architecture to design the web application, and combines the server and browser technology to realize the operation of data interaction with the SQL database. A rectangular bandpass filter is selected for narrow-band information extraction by introducing the sinc function to improve the feature extraction efficiency of speech signals at each frequency. CTC is used to obtain input feature-to-state mapping to solve the sequence problem in speech recognition and achieve end-to-end speech recognition. To achieve the final audio recognition results, the speech decoding algorithm is employed to select the optimal speech output. The results show that in the listening score, the score of the experimental group is 9, and the score of the control is 6. The completion rate of the topics among students who studied on the English teaching platform was higher than that among students who were taught in ordinary English. The completion rate is 0.82 for English questions of grammar type and 0.98 for English questions of listening type, and the average number of passes is the least number of passes, and the number of times is 2.

Применение вероятностных сетей к проблеме распознавания речи

<p>This article discusses an approach to solving the problem of speech recognition based on probabilistic neural networks. The problem is formulated as a problem of command recognition. Commands have equal lengths (in words). Each word position has its own set of candidates. The recognition algorithm for solving this problem was developed. The core of the algorithm is a probabilistic network that recognizes modified estimates of spectral densities. The algorithm allows for high precision of recognition, which is sufficient for the creation of a voice user interface.</p>

Generalization Analysis of CNNs for Classification on Spheres.

Deep learning based on deep convolutional neural networks (CNNs) is extremely efficient in solving classification problems in speech recognition, computer vision, and many other fields. But there is no enough theoretical understanding about this topic, especially the generalization ability of the induced CNN algorithms. In this article, we develop some generalization analysis of a deep CNN algorithm for binary classification with data on spheres. An essential property of the classification problem is the lack of continuity or high smoothness of the target function associated with a convex loss function such as the hinge loss. This motivates us to consider the approximation of functions in the Lp space with 1≤ p ≤ ∞. We provide rates of Lp -approximation when the approximated function lies in a Sobolev space and then present generalization bounds and learning rates for the excess misclassification error of the deep CNN classification algorithm. Our novel analysis is based on efficient cubature formulae on spheres and other tools from spherical analysis and approximation theory.

Research on word stress in Iranian languages by Soviet and Russian scholars

In recent years, considerable material has been accumulated in the field of experimental studies of Iranian languages, including the works by Soviet and Russian scholars, enabling us to make new generalizations regarding the acoustic characteristics of word stress as part of the problem of speech recognition. The study of Iranian languages has been rather uneven: most of the acoustic studies focused on Persian, and only a few covered other 11 languages described in this article. In addition, most of these studies have been published in Russian and therefore remain unknown to the wide international linguistic community. The purpose of the article is to sum up the achievements of Soviet and Russian scholars regarding the acoustic properties of the stressed syllable in Iranian languages. Different views of Soviet, Russian and foreign authors were compared. A number of positions with weak points in reasoning were screened out, and the most well-reasoned ones adopted as the most probable traits of word stress in Iranian languages. Tonal stress was found in Mazandarani, Persian and Tajik; quantitative - in Dari (Afghanistan), Sarikoli and theoretically in Rushani; multicomponental - in Abyanei, Gavruni, Gilaki, Pashto, and Wakhi. Ossetic is likely to have quantitative stress, but statistical proof hasn’t been found yet. Apparently, the overall situation reveals that tonal and quantitative stress types are typical for many Iranian languages. Dynamic stress is found in several languages, but only as a part of multicomponental one; and spectral stress is the rarest feature. The results achieved could be used in automated transcription and speech recognition services.

Application of virtual human sign language translation based on speech recognition

For the application problem of speech recognition to sign language translation, we conducted a study in two parts: improving speech recognition's effectiveness and promoting the application of sign language translation. The mainstream frequency-domain feature has achieved great success in speech recognition. However, it fails to capture the instantaneous gap in speech, and the time-domain feature makes up for this deficiency. In order to combine the advantages of frequency and time domain features, an acoustic architecture with a joint time domain encoder and frequency domain encoder is proposed. A new time-domain feature based on SSM (State-Space-Model) is proposed in the time- domain encoder and encoded using the GRU model. A new model, ConFLASH, is proposed in the frequency domain encoder, which is a lightweight model combining CNN and FLASH (a variant of the Transformer model). It not only reduces the computational complexity of the Transformer model but also effectively integrates the global modeling advantages of the Transformer model and the local modeling advantages of CNN. The Transducer structure is used to decode speech after the encoders are joined. This acoustic model is named GRU-ConFLASH- Transducer. On the self-built dataset and open-source dataset speechocean, it achieves optimal WER (Word Error Rate) of 2.6% and 4.7%. In addition, to better realize the visual application of sign language translation, a 3D virtual human model is designed and developed.

A Data-Centric Approach for Portuguese Speech Recognition: Language Model And Its Implications

Recent advances in Automatic Speech Recognition have made it possible to achieve a quality never seen before in the literature, both for languages with abundant data, such as English, which has a large number of studies and for the Portuguese language, which has a more limited amount of resources and studies. The most recent advances address speech recognition problems with Transformers based models, which have the capability to perform the speech recognition task directly from the raw signal, without the need for manual feature extraction. Some studies have already shown that it is possible to further improve the quality of the transcription of these models using language models within the decoding stage, however, the real impact of such language models is still not clear, especially for the Brazilian Portuguese scenario. Also, it is known that the quality of the data used for training the models is of paramount importance, however, there are few works in the literature addressing this issue. This work explores the impact of language models applied to Portuguese speech recognition both in terms of data quality and computational performance, with a data-centric approach. We propose an approach to measure similarity between datasets and, thus, assist in decision-making during training. The approach indicates paths for the advancement of the state-of-the-art aiming at Portuguese speech recognition, showing that it is possible to reduce the size of the language model by 80% and still achieve error rates around 7.17% for the Common Voice dataset. The source code is available at https://github.com/joaoalvarenga/language-model-evaluation.

The Use of Correlation Features in the Problem of Speech Recognition

The problem solved in the article is connected with the increase in the efficiency of phraseological radio exchange message recognition, which sometimes takes place in conditions of increased tension for the pilot. For high-quality recognition, signal preprocessing methods are needed. The article considers new data preprocessing algorithms used to extract features from a speech message. In this case, two approaches were proposed. The first approach is building autocorrelation functions of messages based on the Fourier transform, the second one uses the idea of building autocorrelation portraits of speech signals. The proposed approaches are quite simple to implement, although they require cyclic operators, since they work with pairs of samples from the original signal. Approbation of the developed method was carried out with the problem of recognizing phraseological radio exchange messages in Russian. The algorithm with preliminary feature extraction provides a gain of 1.7% in recognition accuracy. The use of convolutional neural networks also provides an increase in recognition efficiency. The gain for autocorrelation portraits processing is about 3–4%. Quantization is used to optimize the proposed models. The algorithm’s performance increased by 2.8 times after the quantization. It was also possible to increase accuracy of recognition by 1–2% using digital signal processing algorithms. An important feature of the proposed algorithms is the possibility of generalizing them to arbitrary data with time correlation. The speech message preprocessing algorithms discussed in this article are based on classical digital signal processing algorithms. The idea of constructing autocorrelation portraits based on the time series of a signal has a novelty. At the same time, this approach ensures high recognition accuracy. However, the study also showed that all the algorithms under consideration perform quite poorly under the influence of strong noise.

Speech Audio Modeling by Means of Causal Moving Average Equipped Gated Attention

In the paper we compare different attention mechanisms on the task of audio generation using unsupervised approaches following previous work in language modeling. It is important problem, as far as speech synthesis technology could be used to convert textual information into acoustic waveform signals. These representations can be conveniently integrated into mobile devices and used in such applications as voice messengers or email apps. Sometimes it is difficult to understand and read important messages when being abroad. The lack of appropriate computer systems or some security problems may arise. With this technology, e-mail messages can be listened quickly and efficiently on smartphones, boosting productivity. Apart from that, it is used to assist visually impaired people, so that, for instance, the screen content can be automatically read aloud to a blind user. Nowadays, home appliances, like slow cookers can use this system too for reading culinary recipes, automobiles for voice navigation to the destination spot, or language learners for pronunciation teaching. Speech generation is the opposite problem of automatic speech recognition (ASR) and is researched since the second half of the eighteen's century. Also, this technology also helps vocally handicapped people find a way to communicate with others who do not understand sign language. However, there is a problem, related to the fact that the audio sampling rate is very high, thus leading to very long sequences which are computationally difficult to model. Second challenge is that speech signals with the same semantic meaning can be represented by a lot of signals with significant variability, which is caused by channel environment, pronunciation or speaker timbre characteristics. To overcome these problems, we train an autoencoder model to discretize continuous audio signal into a finite set of discriminative audio tokens which have a lower sampling rate. Subsequently, autoregressive models, which are not conditioned on text, are trained on this representation space to predict the next token, based on previous sequence elements. Hence, this modeling approach resembles causal language modeling. In our study, we show that unlike in the original MEGA work, traditional attention outperforms moving average equipped gated attention, which shows that EMA gated attention is not stable yet and requires careful hyper-parameter optimization.

Computing Topological Invariants of Deep Neural Networks.

A deep neural network has multiple layers to learn more complex patterns and is built to simulate the activity of the human brain. Currently, it provides the best solutions to many problems in image recognition, speech recognition, and natural language processing. The present study deals with the topological properties of deep neural networks. The topological index is a numeric quantity associated to the connectivity of the network and is correlated to the efficiency and accuracy of the output of the network. Different degree-related topological indices such as Zagreb index, Randic index, atom-bond connectivity index, geometric-arithmetic index, forgotten index, multiple Zagreb indices, and hyper-Zagreb index of deep neural network with a finite number of hidden layers are computed in this study.

Automated Speech Audiometry for Integrated Voice Over Internet Protocol Communication Services.

Problems in speech recognition are often apparent in telecommunication situations. For ecologically valid assessments of such conditions, it is important to quantify the impact of real environments including acoustic conditions at a far-end communication device and all paths of transmission degradation. This study presents an automated matrix sentence test procedure based on automatic speech recognition (ASR) integrated in a Voice over Internet Protocol (VoIP) infrastructure and compares the individual effects of transmission degradations with results from laboratory measurements. Speech recognition thresholds (SRTs) were measured in 16 normal-hearing subjects in four test conditions: (a) a laboratory condition guided by a human experimenter, (b) a laboratory condition with reduced bandwidth and (c) additionally reduced headset quality to simulate typical communication systems, and (d) an automated, ASR-controlled adaptive test procedure over a real VoIP infrastructure. Errors of the ASR system were analyzed to show possible effects on measurement outcome Results: Measured SRTs showed a highly significant correlation (r = .93) between the fully automatic and "laboratory" conditions, with a constant bias of about 1 dB indicating a linear shift of the data without affecting the distribution around the mean. The individual impact of the different system degradations on SRTs could be quantified Conclusions: This study provides a proof of concept for automated ASR-based SRT measurements over VoIP systems for speech audiometric testing in real communication systems, as it produced results comparable to traditional laboratory settings for this group of 16 normal-hearing subjects. This makes VoIP services a promising candidate for speech audiometric testing in real communication systems.

Pattern Classification Using Supervised Hypersphere Neural Network0

Pattern classification is of significant demand in the field of machine learning. Its applications range from a simple problem of speech recognition to a complex and important problem of medical diagnosis. Fuzzy based algorithms have been one of the most important methods which have contributed in solving the pattern classification problem. A customized Fuzzy based Supervised Hypersphere Neural Network (SHNN) is presented for the use of pattern classification. Here, a modernized pattern classification method has been presented by considering the fuzzy hypersphere neural network concept at the back end and using a modified version of the membership function aiming to solve the pattern classification problems and boost the performance of the algorithm. The proposed SHNN model creates supervised hypersphere using measurements obtained from intra-class distance techniques along with individual class pattern choice, over the fuzzy membership function. The previous modified fuzzy approaches presented an inherent drawback of ambiguous assignment of classes, losing the fuzzy nature during the assigning of classes in the testing phase and over fitting of the model during the training phase. The proposed approach solves this problem by adding a non linearity to the output of the membership function to maintain its fuzzy nature. Additionally, a new weighted Euclidean distance equation has been designed to enhance the performance of the algorithm. The performance of the proposed model of SHNN has been examined on four standard datasets namely - Pima, liver, glass and monks-3. The results obtained were superior to the previous proposed approaches. Thus, presenting a new state result of the art on the datasets.

Multilingual Speech to Text using Deep Learning based on MFCC Features

The proposed methodology presented in the paper deals with solving the problem of multilingual speech recognition. Current text and speech recognition and translation methods have a very low accuracy in translating sentences which contain a mixture of two or more different languages. The paper proposes a novel approach to tackling this problem and highlights some of the drawbacks of current recognition and translation methods. The proposed approach deals with recognition of audio queries which contain a mixture of words in two different languages - Kannada and English. The novelty in the approach presented, is the use of a next Word Prediction model in combination with a Deep Learning speech recognition model to accurately recognise and convert the input audio query to text. Another method proposed to solve the problem of multilingual speech recognition and translation is the use of cosine similarity between the audio features of words for fast and accurate recognition. The dataset used for training and testing the models was generated manually by the authors as there was no pre-existing audio and text dataset which contained sentences in a mixture of both Kannada and English. The DL speech recognition model in combination with the Word Prediction model gives an accuracy of 71% when tested on the in-house multilingual dataset. This method outperforms other existing translation and recognition solutions for the same test set. Multilingual translation and recognition is an important problem to tackle as there is a tendency for people to speak in a mixture of languages. By solving this problem, the barrier of language and communication can be lifted and thus can help people connect better and more comfortably with each other.

Survey on optimization algorithms in speech processing

Nature is a tremendous provenance of resolving hard and complex problems that exist in the field of computer science because it reveals a very diverse, robust, dynamic, and interesting phenomenon. It constantly finds the optimal solution to resolve its problem which accomplishes exact equity among its element. Nature-inspired algorithms are the heuristic high-level procedure that interprets nature to solve the optimization problem which popularized in the new era of computation. The main objective of this paper is to evaluate the modern technology and enhancement in the nature-inspired algorithm, especially in the application of speech processing, speech recognition, and speech feature selection problems. This paper presents broad collections of global optimization algorithms which have been successfully applied to generate recognition systems that are integrated with metaheuristic algorithms.

Effects of Explicit Connected Speech Instruction on Libyan EFL Learners’ Perceptions of Listening Comprehension Problems in Speech Recognition, Parsing and Utilisation

This quasi-experimental study examined the effects of explicit connected speech instruction in reducing EFL learners’ listening comprehension problems associated with speech recognition, parsing and utilization. Sixty-two (N = 62) Libyan intermediate-level EFL learners, aged 19 to 21, served as the subjects. They were randomly assigned to either the treatment (n = 32) or control group (n = 30). The treatment subjects were exposed to explicit instruction on the features of connected speech for one term (i.e., 12 weeks). An adapted Listening Comprehension Processing Problems questionnaire was administered to both groups before and after the treatment in order to investigate the problems they encountered in comprehending connected speech. The questionnaire contained items measuring language comprehension difficulties in three categories, i.e., speech recognition, parsing, and utilisation. Following the 12-week intervention, the treatment group showed a significant improvement in their report of comprehension problems, while the control group remained constant or unchanged in their perception of the problems. The effect sizes of the differences between the treatment and control group were large, ranging between d = 1.14 (parsing) and d =1.19 (recognition). The findings highlight the necessity to teach connected speech recognition in the EFL curriculum to improve learners’ listening comprehension ability.

Identifying the influence of transfer learning method in developing an end-to-end automatic speech recognition system with a low data level

Ensuring the best quality and performance of modern speech technologies, today, is possible based on the widespread use of machine learning methods. The idea of this project is to study and implement an end-to-end system of automatic speech recognition using machine learning methods, as well as to develop new mathematical models and algorithms for solving the problem of automatic speech recognition for agglutinative (Turkic) languages. Many research papers have shown that deep learning methods make it easier to train automatic speech recognition systems that use an end-to-end approach. This method can also train an automatic speech recognition system directly, that is, without manual work with raw signals. Despite the good recognition quality, this model has some drawbacks. These disadvantages are based on the need for a large amount of data for training. This is a serious problem for low-data languages, especially Turkic languages such as Kazakh and Azerbaijani. To solve this problem, various methods are needed to apply. Some methods are used for end-to-end speech recognition of languages belonging to the group of languages of the same family (agglutinative languages). Method for low-resource languages is transfer learning, and for large resources – multi-task learning. To increase efficiency and quickly solve the problem associated with a limited resource, transfer learning was used for the end-to-end model. The transfer learning method helped to fit a model trained on the Kazakh dataset to the Azerbaijani dataset. Thereby, two language corpora were trained simultaneously. Conducted experiments with two corpora show that transfer learning can reduce the symbol error rate, phoneme error rate (PER), by 14.23 % compared to baseline models (DNN+HMM, WaveNet, and CNC+LM). Therefore, the realized model with the transfer method can be used to recognize other low-resource languages.

A Deep Convolutional Neural Network Based Approach for Effective Neonatal Cry Classification

Background: The cry is the universal language for babies to communicate with others. Infant cry classification is a kind of speech recognition problem that should be treated wisely. In the last few years, it has been gaining its momentum which will be very helpful for the caretaker. Objective: This study aims to develop infant cry classification system predictive model by converting the audio signals into spectrogram image then implementing deep convolutional neural network. It performs end to end learning process and thereby reducing the complexity involved in audio signal analysis and improves the performance using optimization technique. Method: A time frequency-based analysis called Short Time Fourier Transform (STFT) is applied to generate the spectrogram. 256 DFT (Discrete Fourier Transform) points are considered to compute the Fourier transform. A Deep convolutional neural network called AlexNet with few enhancements is done in this work to classify the recorded infant cry. To improve the effectiveness of the above mentioned neural network, Stochastic Gradient Descent with Momentum (SGDM) is used to train the algorithm. Results: A deep neural network-based infant cry classification system achieves a maximum accuracy of 95% in the classification of sleepy cries. The result shows that convolutional neural network with SGDM optimization acquires higher prediction accuracy. Conclusion: Since this proposed work is compared with convolutional neural network with SGD and Naïve Bayes and based on the result, it is implied the convolutional neural network with SGDM performs better than the other techniques.

Directional Adversarial Training for Robust Ownership-Based Recommendation System

Machine learning algorithms are susceptible to cyberattacks, posing security problems in computer vision, speech recognition, and recommendation systems. So far, researchers have made great strides in adopting adversarial training as a defensive strategy. Single-step adversarial training methods have been proposed as viable solutions for improving model generality and resilience. However, there has been little study to address this issue in the context of ownership-based recommendations, which may fail to capture stable results. In this work, we adapt the single-step adversarial training for ownership recommendation systems. Our main technical contributions are as follows: (1) We propose Adversarial Consumption and Production Relationship (ACPR), a model that combines factorization machine and single-step adversarial training for ownership recommendations. It enables us to take advantage of modeling consumption-production interactions with a factorization machine instead of the conventional matrix factorization method for ownership recommendations. (2) We enrich the ACPR technique with directional adversarial training and call our technique Adversarial Consumption and Production Relationship-Aware Directional Adversarial Model (ACPR-ADAM). The idea behind our ACPR-ADAM is that instead of the worst perturbation direction, the perturbation direction in the embedding space is restricted to other examples in the current embedding space, allowing us to incorporate the collaborative signal into the training process. Lastly, through extensive evaluations on Reddit and Pinterest, we demonstrate that our proposed method outperforms state-of-the-art methods. Compared with CPR and ACPR on Reddit and Pinterest datasets, our proposed ACPR-ADAM achieves 93%, 88%, and 72%, 69% enhancement in terms of AUC and HR, respectively.