Voice Assistant Applications Research Articles

Scalable Transformer Accelerator with Variable Systolic Array for Multiple Models in Voice Assistant Applications

Transformer model is a type of deep learning model that has quickly become fundamental in natural language processing (NLP) and other machine learning tasks. Transformer hardware accelerators are usually designed for specific models, such as Bidirectional Encoder Representations from Transformers (BERT), and vision Transformer models, like the ViT. In this study, we propose a Scalable Transformer Accelerator Unit (STAU) for multiple models, enabling efficient handling of various Transformer models used in voice assistant applications. Variable Systolic Array (VSA) centralized design, along with control and data preprocessing in embedded processors, enables matrix operations of varying sizes. In addition, we propose an efficient variable structure and a row-wise data input method for natural language processing where the word count changes. The proposed scalable Transformer accelerator accelerates text summarization, audio processing, image search, and generative AI used in voice assistance.

Enhancing Voice Assistants with Generative AI Language Models for Interactive Conversations

This paper presents comprehensive research for the development of a voice assistant application using HTML/CSS, JavaScript, Python, SQLite, and the Hugging Face API. The proposed methodology encompasses key steps in data collection, model training, frontend and backend development, integration, testing, and deployment, ensuring the creation of a robust and user-friendly voice assistant system. By leveraging a combination of web technologies and Al tools, the developed voice assistant application offers an interactive and intelligent user experience across its users. Key Words: Voice assistant, HTML/CSS, JavaScript, Python, SQLite, Hugging Face API, LLMs, AI.

Application of building BIM technology based on sensor systems and 5G cloud computing

At present, the technology of artificial intelligence has been developing, making people's life more convenient and production more efficient. As the artificial intelligence software Siri is accepted and used by more and more people, some voice assistant software with similar functions gradually appear in people's vision, such as Sogou voice assistant, Xiaoai, Xiaodu and so on. ASR is the entrance of Siri and other software systems. The quality of software is directly affected by ASR. If there is no good ASR, it is impossible to have high-quality voice assistant applications. Therefore, we must correctly understand the important value of ASR and deeply study this technology in order to achieve better human-computer interaction. In recent years, the development of communication network technology is changing rapidly, from 2G to 5G, technology has been in progress. Compared with the previous 4G, 5G has a higher technical level, whether in network equipment or in data computing technology. For 5 g network, cloud computing technology is very important and has a great impact on it, so we need to focus on the research to realize the optimization and innovation development of 5G network. The construction sector constitutes a crucial segment of China's economy, exerting a significant influence on its overall growth trajectory. Given the prevailing developmental context, this document provides a concise overview of the BIM (Building Information Modeling) Technology, encompassing its fundamental concept, distinguishing attributes, and core principles. Furthermore, it delves into the manifold junctures where BIM Technology finds application across the entire construction life cycle. By employing an actual project as an illustrative instance, this study compiles a synopsis of BIM Technology's utilization at each phase, conducting a comprehensive evaluation of the project through the lens of the accrued experience and outcomes. This assessment lays the groundwork for an extensive exploration of potential expansion and application avenues, culminating in the eventual advancement of BIM Technology's achievements.

Acoustic correlates of perceived personality from Korean utterances in a formal communicative setting.

The aim of the present study was to find acoustic correlates of perceived personality from the speech produced in a formal communicative setting-that of Korean customer service employees in particular. This work extended previous research on voice personality impressions to a different sociocultural and linguistic context in which speakers are expected to speak politely in a formal register. To use naturally produced speech rather than read speech, we devised a new method that successfully elicited spontaneous speech from speakers who were role-playing as customer service employees, while controlling for the words and sentence structures they used. We then examined a wide range of acoustic properties in the utterances, including voice quality and global acoustic and segmental properties using Principal Component Analysis. Subjects of the personality rating task listened to the utterances and rated perceived personality in terms of the Big-Five personality traits. While replicating some previous findings, we discovered several acoustic variables that exclusively accounted for the personality judgments of female speakers; a more modal voice quality increased perceived conscientiousness and neuroticism, and less dispersed formants reflecting a larger body size increased the perceived levels of extraversion and openness. These biases in personality perception likely reflect gender and occupation-related stereotypes that exist in South Korea. Our findings can also serve as a basis for developing and evaluating synthetic speech for Voice Assistant applications in future studies.

Security and privacy problems in voice assistant applications: A survey

Voice assistant applications have become omniscient nowadays. Two models that provide the two most important functions for real-life applications (i.e., Google Home, Amazon Alexa, Siri, etc.) are Automatic Speech Recognition (ASR) models and Speaker Identification (SI) models. According to recent studies, security and privacy threats have also emerged with the rapid development of the Internet of Things (IoT). The security issues researched include attack techniques toward machine learning models and other hardware components widely used in voice assistant applications. The privacy issues include technical-wise information stealing and policy-wise privacy breaches. The voice assistant application takes a steadily growing market share every year, but their privacy and security issues never stopped causing huge economic losses and endangering users' personal sensitive information. Thus, it is important to have a comprehensive survey to outline the categorization of the current research regarding the security and privacy problems of voice assistant applications. This paper concludes and assesses five kinds of security attacks and three types of privacy threats in the papers published in the top-tier conferences of cyber security and voice domain.

The communicative potential of Russian-speaking and English-speaking voice systems

The article describes the analysis of the "spontaneous" dialogue of two Russian and foreign voice systems. The study is relevant because the possibilities of communication of the introduction spontaneously dialogue between Russian-speaking and foreign voice systems have not been studied by now. The objects of research have become two Russian-speaking systems – Alice by Yandex, voice assistant application by Sberbank; two English voice systems – Siri by Apple Ink and Google Assistant, including their Russian version. The choice is due to their popularity among native speakers of the Russian language, which was determined by the requests of these voice assistants on the Internet. The sampling period was from January 1, 2021 to December 20, 2022. Russian-speaking and English-speaking voice systems were selected by a continuous sampling method. The article draws conclusions, including about the communicative capabilities of Russian-speaking voice systems and foreign analogues with a built-in translation function into Russian. The scientific novelty is the analysis of "spontaneous" dialogic speech and the analysis of its compliance with the norms of the modern Russian language. At the end of the article there are results of the analysis of foreign phrases and expressions with a small number of variants are summarized to generate a response to the user based on materials from the corpus of the Russian language. Also, it is not uncommon to identify features of machine translation in foreign voice analogues, unlike Russian-speaking voice assistants.

DetBERT: Enhancing Detection of Policy Violations for Voice Assistant Applications using BERT

DetBERT: Enhancing Detection of Policy Violations for Voice Assistant Applications using BERT

Performance Improvement on Traditional Chinese Task-Oriented Dialogue Systems With Reinforcement Learning and Regularized Dropout Technique

The development of conversational voice assistant applications has been in full swing around the world. This paper aims to develop traditional Chinese multi-domain task-oriented dialogue (TOD) systems. It is typically implemented using pipeline approach, where submodules are optimized independently, resulting in inconsistencies with each other. Instead, this paper implements end-to-end multi-domain TOD models using pre-trained deep neural networks (DNNs). This allows us to integrate all the submodules into one single DNN model to solve the inconsistencies. Data shortages are common in conversational natural language processing (NLP) tasks using DNN models. In this regard, dropout regularization has been widely used to improve overfitting caused by insufficient training dataset. However, the randomness it introduces leads to non-negligible discrepancies between training and inference. On the other hand, pre-trained language models have successfully provided effective regularization for NLP tasks. An inherent disadvantage is that fine-tuning the pre-trained language model suffers from exposure bias and loss-evaluation mismatch. To this end, we propose a reinforcement learning (RL) approach to address both issues. Furthermore, we adopt a method called regularized dropout (R-Drop) to improve the inconsistency in dropout layers of DNNs. Experimental results show that both our proposed RL approach and the R-Drop technique can significantly improve the joint target accuracy (JGA) score and combined score of traditional Chinese TOD system in tasks of dialogue state tracking (DST) and end-to-end sentence prediction, respectively.

The Effectiveness of Using Mobile Interactive Voice Assistant Applications in Developing Academic Self-Efficacy of Saudi University Students during the COVID-19 PandemicThe Effectiveness of Using Mobile Interactive Voice Assistant Applications in Developing Academic Self-Efficacy of Saudi University Students during the COVID-19 Pandemic

The Effectiveness of Using Mobile Interactive Voice Assistant Applications in Developing Academic Self-Efficacy of Saudi University Students during the COVID-19 PandemicThe Effectiveness of Using Mobile Interactive Voice Assistant Applications in Developing Academic Self-Efficacy of Saudi University Students during the COVID-19 Pandemic

Effects of voice assistant creation using different learning approaches on performance of computational thinking

Designing artificial intelligence (AI) artefact learning has gone beyond command-line-based instruction, to include a low-barrier threshold with block-based programming. Such instructional design must not solely emphasise AI workings. Rather, it must offer students computational thinking (CT) practice to support their AI-related artefact creation while reducing their AI anxiety about future job replacement or sociotechnical blindness. In this study, this research explored an experiential learning approach to improve CT along with AI application capabilities when engaging undergraduate students in creating a voice assistant application (VA app). A total of 56 students participated in the study. The control group (CG) of 26 students used a conventional subject-based learning method, while the experimental group (EG) of 30 students adopted an experiential learning method. This study aimed to examine the differences in the learning achievement of CT and AI concept, as well as the perspectives of AI anxiety, and CT; in the meanwhile, this study analysed the students' learning behaviours using sequential behavioural analysis to discuss the learning process. Results showed that the CT ability of the EG was better than that of the CG, although no significant difference was found between the two groups’ AI concepts and anxiety. The behaviour analysis also revealed that the EG students were willing to ask more questions, and conducted their VA evaluation, whereas the CG students were inclined to focus on the input and output of knowledge, and replicated what the teacher presented. Suggestions and implications are given for future research.

Problematic Privacy Policies of Voice Assistant Applications

This work conducts a comprehensive empirical analysis on the privacy policies of 64,720 Amazon Alexa skills and 16,002 Google Assistant actions. We show that a substantial number of problematic privacy policies exist in the Amazon Alexa and Google Assistant platforms.

Virtual voice assistant applications improved expressive verbal abilities and social interactions in children with autism spectrum disorder: a Single-Subject experimental study

The use of interactive technologies has been demonstrated to enhance verbal and non-verbal communication, as well as the social interaction tendencies of children with Autism Spectrum Disorder (ASD). We examined effects of using Virtual Voice Assistant (VVAs) in children with ASD with respect to two outcomes: speech skills and social interaction skills. A single-case study included three children with ASD (4–11 years old) that utilized VVAs for three months. Pre- and post-intervention questionnaires and semi-structured interviews were used to measure the communication and social interaction skills of the participating children. Participant One, Two and Three showed improvement in the number of correct words produced the VVA intervention. All participants showed increases in social interactions in the intervention phase. Overall, the results showed that the VVAs had positive effects on the speech and social interaction skills of autistic children. The findings demonstrate that children with ASD may benefit from VVAs to improve their communication skills.

Analysing antecedence of an intelligent voice assistant use intention and behaviour

Background: As the use of intelligent voice assistant applications becomes more prevalent, a growing body of studies are examining individuals' interactions with intelligent voice assistants. However, very limited research has focused on comparing the antecedents of both use and non-use behaviour of individuals, based on the technology acceptance models. To fill this gap, the present study investigated antecedents of intelligent voice assistance use and use intention in a cross-sectional setting. Additionally, to go one step beyond the existing literature on technology acceptance models and theories, a new construct termed perceived needs, as well as the moderating role of perceived privacy concerns and perceived awareness, are introduced. Method: A quantitative, cross-sectional research design was utilised using a nonprobability sampling strategy through the online networking platforms. Total of 277 (n = 155 users vs n = 122 non-users) international adults age between 20-74 years (79.6% female, 20.4 % male) contributed to the study. Ordinary least squares (OLS) linear regression and Bivariate logistic regression analyses for non-users and users were conducted, respectively. Results: Both analyses revealed that peer influence and perceived needs related to the intention to use intelligent voice assistants for non-users, which applied to the current intelligent voice assistance users where privacy concerns were considered. Surprisingly, the key determinants of technology acceptance and use theories, such as perceived ease of use and perceived usefulness, did not hold for intelligent voice assistance usage. Conclusion: The current research contributed to the field by validating new constructs of perceived needs and the moderation role of perceived privacy concerns. However, in order to build on an existing body of knowledge, future studies should further examine the moderation role of perceived privacy concerns, perceived ease of use, and perceived usefulness in the same domain.

Voice Assistant Application for Avoiding Sedentarism in Elderly People Based on IoT Technologies

The rise in the use of virtual assistants such as Siri, Google Assistant, or Alexa among different sectors of society is facilitating access to information and services that were previously inconceivable due to the existing digital divide due to age. This situation allows especially the elderly to perform tasks much more easily and to access applications and services that could be a challenge for them with other digital user interfaces. With this in mind, the EMERITI project aims to improve the lives of the elderly through the use of virtual assistants in different case studies. In this sense, virtual voice assistants along with the use of Internet of Things (IoT) technologies can contribute to avoid sedentarism in the elderly; however, it is necessary to address the problem of proactivity presented by the virtual assistants available in the market. This article presents a solution that, through the use of activity monitoring smart bracelets, IoT devices and virtual voice assistants allow the elderly to monitor their daily physical activity simply by using their voice and therefore prevent them from sedentary patterns. Finally, this study presents the technical results obtained after the deployment of the proposed system and discusses the main advantages and the current challenges of the use of virtual assistants in applications to prevent sedentary lifestyles in the elderly.

Semantic voice search in IETP: filling the gap for maintenance 4.0

PurposeA framework is being developed to help Integrated Electronic Technical Publications (IETP) consultation inside and outside the aviation maintenance hangar. The expected results are the reduction in time to access the desired IETP and to assist mechanics while performing maintenance tasks using voice recognition.Design/methodology/approachThe work is being conducted based on literature review and consultation with mechanics from the aviation industry, through questionnaires. The development will be made through study cases by building a core search engine and mobile applications to support the mechanics during the maintenance activities.FindingsThe identified problem in small maintenance shops and defence organizations suggests that IETP are not entirely accessible before and during the maintenance activity. Such organizations suffer from information and communications technology (ICT) low infrastructure capability and demand access to multiple IETP databases as they usually support different aircraft. To have access to the IETP through voice assistant application will help mechanics to access the IETP, including when they would be with dirty hands and having difficulty in using mobile devices with touch displays.Originality/valueThe framework being developed will give mechanics the ability to quickly find any existing IETP to support its maintenance task at any time and in any place with low demanding for ICT infrastructure. The architecture will support different applications, and the identified priority is for IETP viewers to the most demanding functionality of specification ASD S1000D. This approach could also help in troubleshooting activities since COVID-19 brought new demands for the social distancing for mechanics.

Mobile applications aiming to facilitate immigrants’ societal integration and overall level of integration, health and mental health. Does artificial intelligence enhance outcomes?

Using panel data on immigrant populations from European, Asian and African countries the study estimates positive associations between the number of mobile applications in use aiming to facilitate immigrants' societal integration (m-Integration) and increased level of integration (Ethnosizer), good overall health (EQ-VAS) and mental health (CESD-20). It is estimated that the patterns are gender sensitive. In addition, it is found that m-Integration applications in relation to translation and voice assistants, public services, and medical services provide the highest returns on immigrants' level of integration, health/mental health status. For instance, translation and voice assistant applications are associated with a 4% increase in integration and a 0.8% increase in good overall health. Moreover, m-Integration applications aided by artificial intelligence (AI) are associated with increased health/mental health and integration levels among immigrants. We indicate that AI by providing customized search results, peer reviewed e-learning, professional coaching on pronunciation, real-time translations, and virtual communication for finding possible explanations for health conditions might bring better quality services facilitating immigrants' needs. This is the first known study to introduce the term ‘m-Integration’, quantify associations between applications, health/mental health and integration for immigrants, and assess AI's role in enhancing the aforementioned outcomes.