Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.04
Andrei COZMA, Nicolae ȚĂPUȘ
This paper offers an in-depth exploration of the verification procedure pertaining to the integration of an Advanced High-performance Bus (AHB) Interconnect module. The verification process holds pivotal significance within chip development, entailing thorough validation and examination of the hardware design before entering the mass production phase. The principal objective of this verification process is to meticulously unearth potential bugs or imperfections embedded in the design, which could potentially trigger undesirable outcomes, compromised performance, or even critical malfunctions in the final product. At the heart of this verification approach lies the functional verification paradigm, centered around simulation-based testing. Within this framework, the AHB Interconnect module is instantiated in a controlled verification environment designed to emulate real-world scenarios. This environment orchestrates input stimuli to the module and captures ensuing outputs generated by it. The environment is meticulously programmed to anticipate specific behavioral patterns from the module. Deviations from these anticipated behaviors are promptly flagged as errors. This meticulous methodology serves to guarantee that the module aligns with its intended operations and strictly adheres to predefined functional benchmarks. The paper is dedicated to Acad. Florin Gheorghe Filip, at his 15th anniversary as the Chairman of the Information Science and Technology Section of the Romanian Academy, and at his 75th anniversary.
{"title":"AHB Interconnect - Functional Verification","authors":"Andrei COZMA, Nicolae ȚĂPUȘ","doi":"10.59277/romjist.2023.3-4.04","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.04","url":null,"abstract":"This paper offers an in-depth exploration of the verification procedure pertaining to the integration of an Advanced High-performance Bus (AHB) Interconnect module. The verification process holds pivotal significance within chip development, entailing thorough validation and examination of the hardware design before entering the mass production phase. The principal objective of this verification process is to meticulously unearth potential bugs or imperfections embedded in the design, which could potentially trigger undesirable outcomes, compromised performance, or even critical malfunctions in the final product. At the heart of this verification approach lies the functional verification paradigm, centered around simulation-based testing. Within this framework, the AHB Interconnect module is instantiated in a controlled verification environment designed to emulate real-world scenarios. This environment orchestrates input stimuli to the module and captures ensuing outputs generated by it. The environment is meticulously programmed to anticipate specific behavioral patterns from the module. Deviations from these anticipated behaviors are promptly flagged as errors. This meticulous methodology serves to guarantee that the module aligns with its intended operations and strictly adheres to predefined functional benchmarks. The paper is dedicated to Acad. Florin Gheorghe Filip, at his 15th anniversary as the Chairman of the Information Science and Technology Section of the Romanian Academy, and at his 75th anniversary.","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135470849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.06
Danijela PROTIC, Miomir STANKOVIC
Anomaly-based intrusion detection systems are designed to scan computer network traffic for abnormal behavior. Binary classifiers based on supervised machine learning have proven to be highly accurate tools for classifying instances as normal or abnormal. Main disadvantages of supervised machine learning are the long processing time and large amount of training data required to ensure accurate results. Two preprocessing steps to reduce data sets are feature selection and feature scaling. In this article, we present a new hyperbolic tangent feature scaling approach based on the linearization of the tangent hyperbolic function and the damping strategy of the Levenberg-Marquardt algorithm. Experiments performed on the Kyoto 2006+ dataset used four high-precision binary classifiers: weighted k-nearest neighbors, decision tree, feedforward neural networks, and support vector machine. It is shown that hyperbolic tangent scaling reduces processing time by more than twofold. An XOR-based detector is proposed to determine conflicting decisions about anomalies. The decisions of the FNN and wk-NN models are compared. It is shown that decisions sometimes turn out differently. The percentage of the opposite decisions has been shown to vary and is not affected by dataset size.
{"title":"XOR-Based Detector of Different Decisions on Anomalies in the Computer Network Traffic","authors":"Danijela PROTIC, Miomir STANKOVIC","doi":"10.59277/romjist.2023.3-4.06","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.06","url":null,"abstract":"Anomaly-based intrusion detection systems are designed to scan computer network traffic for abnormal behavior. Binary classifiers based on supervised machine learning have proven to be highly accurate tools for classifying instances as normal or abnormal. Main disadvantages of supervised machine learning are the long processing time and large amount of training data required to ensure accurate results. Two preprocessing steps to reduce data sets are feature selection and feature scaling. In this article, we present a new hyperbolic tangent feature scaling approach based on the linearization of the tangent hyperbolic function and the damping strategy of the Levenberg-Marquardt algorithm. Experiments performed on the Kyoto 2006+ dataset used four high-precision binary classifiers: weighted k-nearest neighbors, decision tree, feedforward neural networks, and support vector machine. It is shown that hyperbolic tangent scaling reduces processing time by more than twofold. An XOR-based detector is proposed to determine conflicting decisions about anomalies. The decisions of the FNN and wk-NN models are compared. It is shown that decisions sometimes turn out differently. The percentage of the opposite decisions has been shown to vary and is not affected by dataset size.","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135469648","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.02
Gheorghe DUCA, Sergey TRAVIN, Inga ZINICOVSCAIA
This paper, dedicated to Acad. Florin Gheorghe Filip, at his 15th anniversary as the Chairman of the Information Science and Technology Section of the Romanian Academy, suggests an approach to describe the data obtained in biomonitoring studies using mosses, on the example of the Republic of Moldova. In total, 33 moss samples were collected on the territory of Moldova, the elemental composition of which was determined by neutron activation analysis and atomic absorption spectrometry. At the first stage of the work, a correlation analysis was carried out with the ranking of data in the order of decreasing total correlation, which made it possible to preliminarily reduce the number of elements to two factors. At the stage of data sorting, iron, the content of which in the environment can be associated with anthropogenic activity, was chosen as the element determining the rank. The next stage of work was data smoothing using a discrete cosine transform, for which the codes were rewritten and the algorithm was ported to the Excel-VBA environment, which is most suitable for preprocessing and graphical display of experimental data. Also, an algorithm was developed for determining the number of linearly independent (basis) vectors in which the matrix itself can be decomposed. Two principal roots were identified, of which the larger one is several thousand units, and three lower roots, which are less than unit were excluded, since their absolute values differ by two to three or more orders of magnitude downward from the eigenvalues of the influencing components. The paper is dedicated to Acad. Florin Gheorghe Filip, at his 15th anniversary as the Chairman of the Information Science and Technology Section of the Romanian Academy, and at his 75th anniversary.
{"title":"Approach to Evaluate the Data of Moss Biomonitoring Studies: Preprocessing and Preliminary Ranking","authors":"Gheorghe DUCA, Sergey TRAVIN, Inga ZINICOVSCAIA","doi":"10.59277/romjist.2023.3-4.02","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.02","url":null,"abstract":"This paper, dedicated to Acad. Florin Gheorghe Filip, at his 15th anniversary as the Chairman of the Information Science and Technology Section of the Romanian Academy, suggests an approach to describe the data obtained in biomonitoring studies using mosses, on the example of the Republic of Moldova. In total, 33 moss samples were collected on the territory of Moldova, the elemental composition of which was determined by neutron activation analysis and atomic absorption spectrometry. At the first stage of the work, a correlation analysis was carried out with the ranking of data in the order of decreasing total correlation, which made it possible to preliminarily reduce the number of elements to two factors. At the stage of data sorting, iron, the content of which in the environment can be associated with anthropogenic activity, was chosen as the element determining the rank. The next stage of work was data smoothing using a discrete cosine transform, for which the codes were rewritten and the algorithm was ported to the Excel-VBA environment, which is most suitable for preprocessing and graphical display of experimental data. Also, an algorithm was developed for determining the number of linearly independent (basis) vectors in which the matrix itself can be decomposed. Two principal roots were identified, of which the larger one is several thousand units, and three lower roots, which are less than unit were excluded, since their absolute values differ by two to three or more orders of magnitude downward from the eigenvalues of the influencing components. The paper is dedicated to Acad. Florin Gheorghe Filip, at his 15th anniversary as the Chairman of the Information Science and Technology Section of the Romanian Academy, and at his 75th anniversary.","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135469795","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.07
T. T. MIRNALINEE, J. BHUVANA, SB. VINODHINISRI
"In the digital era preserving the privacy of confidential information is the crucial issue in authentication and secret communication. Cryptography, steganography, data anonymization are few techniques used to preserve the information. Steganography is the process of hiding information into other non-secret sources of information like text, image or audio, so that secret information is not visible to the naked eye. This paper proposes a novel Prime Number Based Embedding (PNBE) approach for efficient data hiding to encode the data in images without any visual distortion. This approach proposes a novel scheme for identifying the embedding location of the data in the image and PNBE algorithm for the encoding and decoding of messages. The proposed approach is demonstrated with benchmark images. The data encoding capacity and quality are tested on different sized images. Experimental results show better hiding capacity in terms of bytes ranging from 6,327 to 12,448 based on the size of the image. The quality of the image is described as Peak to Signal Noise Ratio (PSNR), which computes the quality difference between original and compressed images. The efficiency of the proposed approach is demonstrated using PSNR with higher values. It is observed from the results that the proposed approach outperforms the existing ones in terms of embedding capacity and quality of the images. "
{"title":"A Data Hiding Approach for Secured Data Communication","authors":"T. T. MIRNALINEE, J. BHUVANA, SB. VINODHINISRI","doi":"10.59277/romjist.2023.3-4.07","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.07","url":null,"abstract":"\"In the digital era preserving the privacy of confidential information is the crucial issue in authentication and secret communication. Cryptography, steganography, data anonymization are few techniques used to preserve the information. Steganography is the process of hiding information into other non-secret sources of information like text, image or audio, so that secret information is not visible to the naked eye. This paper proposes a novel Prime Number Based Embedding (PNBE) approach for efficient data hiding to encode the data in images without any visual distortion. This approach proposes a novel scheme for identifying the embedding location of the data in the image and PNBE algorithm for the encoding and decoding of messages. The proposed approach is demonstrated with benchmark images. The data encoding capacity and quality are tested on different sized images. Experimental results show better hiding capacity in terms of bytes ranging from 6,327 to 12,448 based on the size of the image. The quality of the image is described as Peak to Signal Noise Ratio (PSNR), which computes the quality difference between original and compressed images. The efficiency of the proposed approach is demonstrated using PSNR with higher values. It is observed from the results that the proposed approach outperforms the existing ones in terms of embedding capacity and quality of the images. \"","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135469796","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.05
Horia-Nicolai TEODORESCU, Mironela PIRNAU
The paper pertains to the field of opinion mining on social networks in view of decision-making. It was designed to determine the effect of the energy crises on the socio-ethical and ethical worries and concerns related to the role of Information and Communication Technology (ICT) in the crisis, as reflected on Twitter. Knowing public opinions during crises is a key component of decision-making. Other purposes include the determination of the perception of the correlations between these categories of concerns and the determination of specificities of distributions of such perceptions, as represented by words. A mixture of tools from lexical, topical, and semantical analysis were applied for generating the bag of words and for analyzing a set of tweets collected during the initial stage of an energy crisis, with special emphasis to coherence indices. The results confirm that energy crises may be associated with such concerns, as expected, but the association probability is low for the 2022 energy crisis, indicating at least a partial success of the mitigation decisions of the respective authorities. Some surprising relationships also occurred from the study. The paper is dedicated to Acad. Florin Gheorghe Filip, at the 75th anniversary.
{"title":"Twitter's Mirroring of the 2022 Energy Crisis: What It Teaches Decision-Makers - A Preliminary Study","authors":"Horia-Nicolai TEODORESCU, Mironela PIRNAU","doi":"10.59277/romjist.2023.3-4.05","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.05","url":null,"abstract":"The paper pertains to the field of opinion mining on social networks in view of decision-making. It was designed to determine the effect of the energy crises on the socio-ethical and ethical worries and concerns related to the role of Information and Communication Technology (ICT) in the crisis, as reflected on Twitter. Knowing public opinions during crises is a key component of decision-making. Other purposes include the determination of the perception of the correlations between these categories of concerns and the determination of specificities of distributions of such perceptions, as represented by words. A mixture of tools from lexical, topical, and semantical analysis were applied for generating the bag of words and for analyzing a set of tweets collected during the initial stage of an energy crisis, with special emphasis to coherence indices. The results confirm that energy crises may be associated with such concerns, as expected, but the association probability is low for the 2022 energy crisis, indicating at least a partial success of the mitigation decisions of the respective authorities. Some surprising relationships also occurred from the study. The paper is dedicated to Acad. Florin Gheorghe Filip, at the 75th anniversary.","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135469650","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.03
Gheorghe ȘTEFAN
"There are many approaches to the concept of information in terms of quantity, but there is a lack of consensus on what information is. We measure it in many ways, but we don’t consistently state what information is. We will try in this text to offer a qualitative image for the concept of information. On this occasion, perhaps we will better understand the role of information at different levels in existence by defining information as a symbolic structure that acts through the meaning it has in a given context."
{"title":"Meaning in Action: a Qualitative Approach of Information","authors":"Gheorghe ȘTEFAN","doi":"10.59277/romjist.2023.3-4.03","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.03","url":null,"abstract":"\"There are many approaches to the concept of information in terms of quantity, but there is a lack of consensus on what information is. We measure it in many ways, but we don’t consistently state what information is. We will try in this text to offer a qualitative image for the concept of information. On this occasion, perhaps we will better understand the role of information at different levels in existence by defining information as a symbolic structure that acts through the meaning it has in a given context.\"","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135470846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.01
Gheorghe PĂUN, Mario J. PÉREZ-JIMÉNEZ, Grzegorz ROZENBERG
We initiate the study of spiking neural P systems associated with infinite sequences, by considering them as computability devices which generate infinite sequences of bits (1 indicates a step when a spike exits the system, and 0 indicates a step when the system does not send a spike to the environment), and as devices which process infinite sequence of bits (for instance, computing Boolean operations or other operations on two input sequences). For both the generating and the transduction case we introduce some basic notions illustrated by numerous examples, establish some basic properties, and formulate a number of research topics.
{"title":"Infinite Spike Trains in Spiking Neural P Systems","authors":"Gheorghe PĂUN, Mario J. PÉREZ-JIMÉNEZ, Grzegorz ROZENBERG","doi":"10.59277/romjist.2023.3-4.01","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.01","url":null,"abstract":"We initiate the study of spiking neural P systems associated with infinite sequences, by considering them as computability devices which generate infinite sequences of bits (1 indicates a step when a spike exits the system, and 0 indicates a step when the system does not send a spike to the environment), and as devices which process infinite sequence of bits (for instance, computing Boolean operations or other operations on two input sequences). For both the generating and the transduction case we introduce some basic notions illustrated by numerous examples, establish some basic properties, and formulate a number of research topics.","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"157 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135470850","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.08
Umit KILIC, Esra SARAC ESSIZ, Mumine KAYA KELES
"Datasets comprise a collection of features; however, not all of these features may be necessary. Feature selection is the process of identifying the most relevant features while eliminating redundant or irrelevant ones. To be effective, feature selection should improve classification performance while reducing the number of features. Existing algorithms can be adapted and modified into feature selectors. In this study, we introduce the implementation of the Anarchic Society Optimization algorithm, a human-inspired algorithm, as a feature selector. This is the first study that utilizes the binary version of the algorithm for feature selection. The proposed Binary Anarchic Society Algorithm is evaluated on nine datasets and compared to three known algorithms: Binary Genetic Algorithm, Binary Particle Swarm Optimization, and Binary Gray Wolf Optimization. Additionally, four traditional feature selection techniques (Info Gain, Gain Ratio, Chi-square, and ReliefF) are incorporated for performance comparison. Our experiments highlight the competitive nature of the proposed method, suggesting its potential as a valuable addition to existing feature selection techniques."
{"title":"Binary Anarchic Society Optimization for Feature Selection","authors":"Umit KILIC, Esra SARAC ESSIZ, Mumine KAYA KELES","doi":"10.59277/romjist.2023.3-4.08","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.08","url":null,"abstract":"\"Datasets comprise a collection of features; however, not all of these features may be necessary. Feature selection is the process of identifying the most relevant features while eliminating redundant or irrelevant ones. To be effective, feature selection should improve classification performance while reducing the number of features. Existing algorithms can be adapted and modified into feature selectors. In this study, we introduce the implementation of the Anarchic Society Optimization algorithm, a human-inspired algorithm, as a feature selector. This is the first study that utilizes the binary version of the algorithm for feature selection. The proposed Binary Anarchic Society Algorithm is evaluated on nine datasets and compared to three known algorithms: Binary Genetic Algorithm, Binary Particle Swarm Optimization, and Binary Gray Wolf Optimization. Additionally, four traditional feature selection techniques (Info Gain, Gain Ratio, Chi-square, and ReliefF) are incorporated for performance comparison. Our experiments highlight the competitive nature of the proposed method, suggesting its potential as a valuable addition to existing feature selection techniques.\"","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135469791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.10
Serban MIHALACHE, Dragos BURILEANU
Speech emotion recognition (SER) is the task of determining the affective content present in speech, a promising research area of great interest in recent years, with important applications especially in the field of forensic speech and law enforcement operations, among others. In this paper, systems based on deep neural networks (DNNs) spanning five levels of complexity are proposed, developed, and tested, including systems leveraging transfer learning (TL) for the top modern image recognition deep learning models, as well as several ensemble classification techniques that lead to significant performance increases. The systems were tested on the most relevant SER datasets: EMODB, CREMAD, and IEMOCAP, in the context of: (i) classification: using the standard full sets of emotion classes, as well as additional negative emotion subsets relevant for forensic speech applications; and (ii) regression: using the continuously valued 2D arousal-valence affect space. The proposed systems achieved state-of-the-art results for the full class subset for EMODB (up to 83% accuracy) and performance comparable to other published research for the full class subsets for CREMAD and IEMOCAP (up to 55% and 62% accuracy). For the class subsets focusing only on negative affective content, the proposed solutions offered top performance vs. previously published state of the art results.
{"title":"Speech Emotion Recognition Using Deep Neural Networks, Transfer Learning, and Ensemble Classification Techniques","authors":"Serban MIHALACHE, Dragos BURILEANU","doi":"10.59277/romjist.2023.3-4.10","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.10","url":null,"abstract":"Speech emotion recognition (SER) is the task of determining the affective content present in speech, a promising research area of great interest in recent years, with important applications especially in the field of forensic speech and law enforcement operations, among others. In this paper, systems based on deep neural networks (DNNs) spanning five levels of complexity are proposed, developed, and tested, including systems leveraging transfer learning (TL) for the top modern image recognition deep learning models, as well as several ensemble classification techniques that lead to significant performance increases. The systems were tested on the most relevant SER datasets: EMODB, CREMAD, and IEMOCAP, in the context of: (i) classification: using the standard full sets of emotion classes, as well as additional negative emotion subsets relevant for forensic speech applications; and (ii) regression: using the continuously valued 2D arousal-valence affect space. The proposed systems achieved state-of-the-art results for the full class subset for EMODB (up to 83% accuracy) and performance comparable to other published research for the full class subsets for CREMAD and IEMOCAP (up to 55% and 62% accuracy). For the class subsets focusing only on negative affective content, the proposed solutions offered top performance vs. previously published state of the art results.","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135469792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-28DOI: 10.59277/romjist.2023.3-4.09
Costin Andrei BRATAN, Ana Voichita TEBEANU, Gabriela BOBES
This paper presents the effects’ analysis produced by the frequent use of swearing from the perspective of irritability. The analysis was carried out with the help of two psychological questionnaires that were completed by the volunteers before and after the inducement of the negative emotions and automatic recognition functions implemented by Convolutional Neural Networks (CNN), applied for the speech signals of two volunteer groups for whom negative emotions were induced. The CNN architecture uses Mel-frequency cepstral coefficients (MFCCs), obtained from the speech signal, and has 87,944 trainable parameters, the outputs of the network being the 8 main classes of emotion detected by the algorithm (1 neutral, 3 positive, and 4 negative). The CNN also gives information about the negative emotion and irritability level. For the volunteers who swore during the experiment, there is an increase of 14% in negative emotion intensity and of 21% for the irritability level than for the volunteers who didn’t swear during the trials. The use of this current research is the understanding that cursing causes a higher level of irritability.
{"title":"Using Swear Words Increases the Irritability – a Study Using AI Algorithms","authors":"Costin Andrei BRATAN, Ana Voichita TEBEANU, Gabriela BOBES","doi":"10.59277/romjist.2023.3-4.09","DOIUrl":"https://doi.org/10.59277/romjist.2023.3-4.09","url":null,"abstract":"This paper presents the effects’ analysis produced by the frequent use of swearing from the perspective of irritability. The analysis was carried out with the help of two psychological questionnaires that were completed by the volunteers before and after the inducement of the negative emotions and automatic recognition functions implemented by Convolutional Neural Networks (CNN), applied for the speech signals of two volunteer groups for whom negative emotions were induced. The CNN architecture uses Mel-frequency cepstral coefficients (MFCCs), obtained from the speech signal, and has 87,944 trainable parameters, the outputs of the network being the 8 main classes of emotion detected by the algorithm (1 neutral, 3 positive, and 4 negative). The CNN also gives information about the negative emotion and irritability level. For the volunteers who swore during the experiment, there is an increase of 14% in negative emotion intensity and of 21% for the irritability level than for the volunteers who didn’t swear during the trials. The use of this current research is the understanding that cursing causes a higher level of irritability.","PeriodicalId":54448,"journal":{"name":"Romanian Journal of Information Science and Technology","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135469793","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：