Prediction of System Usability through Neural NetworksApplied to Conversational User Interfaces

A comparative evaluation of deep convolutional neural network and deep neural network-based land use/land cover classifications of mining regions using fused multi-sensor satellite data

On-demand emotional support is an expensive and elusive societal need that is exacerbated in difficult times — as witnessed during the COVID-19 pandemic. Prior work in affective crowdsourcing has examined ways to overcome technical challenges for providing on-demand emotional support to end users. This can be achieved by training crowd workers to provide thoughtful and engaging on-demand emotional support. Inspired by recent advances in conversational user interface research, we investigate the efficacy of a conversational user interface for training workers to deliver psychological support to users in need. To this end, we conducted a between-subjects experimental study on Prolific, wherein a group of workers (N=200) received training on motivational interviewing via either a conversational interface or a conventional web interface. Our results indicate that training workers in a conversational interface yields both better worker performance and improves their user experience in on-demand stress management tasks.

On-demand emotional support is an expensive and elusive societal need that is exacerbated in difficult times -- as witnessed during the COVID-19 pandemic. Prior work in affective crowdsourcing has examined ways to overcome technical challenges for providing on-demand emotional support to end users. This can be achieved by training crowd workers to provide thoughtful and engaging on-demand emotional support. Inspired by recent advances in conversational user interface research, we investigate the efficacy of a conversational user interface for training workers to deliver psychological support to users in need. To this end, we conducted a between-subjects experimental study on Prolific, wherein a group of workers (N=200) received training on motivational interviewing via either a conversational interface or a conventional web interface. Our results indicate that training workers in a conversational interface yields both better worker performance and improves their user experience in on-demand stress management tasks.

Although various methods have been developed to evaluate conversational interfaces, there has been a lack of methods specifically focusing on evaluating user experience. This paper reviews the understandings of user experience (UX) in conversational interfaces literature and examines the six questionnaires commonly used for evaluating conversational systems in order to assess the potential suitability of these questionnaires to measure different UX dimensions in that context. The method to examine the questionnaires involved developing an assessment framework for main UX dimensions with relevant attributes and coding the items in the questionnaires according to the framework. The results show that (i) the understandings of UX notably differed in literature; (ii) four questionnaires included assessment items, in varying extents, to measure hedonic, aesthetic and pragmatic dimensions of UX; (iii) while the dimension of affect was covered by two questionnaires, playfulness, motivation, and frustration dimensions were covered by one questionnaire only. The largest coverage of UX dimensions has been provided by the Subjective Assessment of Speech System Interfaces (SASSI). We recommend using multiple questionnaires to obtain a more complete measurement of user experience or improve the assessment of a particular UX dimension.RESEARCH HIGHLIGHTSVarying understandings of UX in conversational interfaces literature. A UX assessment framework with UX dimensions and their relevant attributes. Descriptions of the six main questionnaires for evaluating conversational interfaces. A comparison of the six questionnaires based on their coverage of UX dimensions.

Artificial intelligence in interdisciplinary life science and drug discovery research.

Conversational interfaces have become a hot topic during the last years. Major research groups and technology companies have been making huge investments in research into technologies such as Artificial Intelligence, deep neural networks, machine learning, and natural language understanding with the aim of creating intelligent assistants that will enable users to interact with information and services in a natural, conversational way. However, most of the current conversational interfaces use hand-crafting dialog strategies and architectures tightly coupled to the application domain and are not adapted to the specific requirements and preferences of each user. In this paper, we propose a multi-agent architecture to develop user-adapted conversational interfaces. Our proposal considers two types of agents. Expert agents access different knowledge sources, and decision agents coordinate them to provide a coherent response to the user. We describe our proposal and its practical application to develop a conversational interface that provides bus schedule information.

Self-screening for mental health problems is commonly used to detect and assess symptoms, as a first step in diagnosing a problem, and to give recommendations for possible treatment. This study explores the potential of conversational interfaces in providing screening services for mental health care. A chatbot was developed to perform a screening for attention deficit/hyperactivity disorder (ADHD) in adults by including the items from the Adult ADHD Self-Report Scale (ASRS). We compared the conversational chatbot interface responses with reports on the standardised paper-based ASRS, and evaluated the user interaction with the chatbot. The results showed a match between the two modalities in the screening results. Based on interviews with participants and chatlogs we discuss the challenges and user experience of doing self-screening in a conversational interface.

Administrative processes are ubiquitous in modern life and have been identified as a particular burden to those with accessibility needs. Students who have accessibility needs often have to understand guidance, fill in complex forms, and communicate with multiple parties to disclose disabilities and access appropriate support. Conversational user interfaces (CUIs) could allow us to reimagine such processes, yet there is currently limited understanding of how to design these to be accessible, or whether such an approach would be preferred. In the ADMINS (Assistants for the Disclosure and Management of Information about Needs and Support) project, we implemented a virtual assistant (VA) which is designed to enable students to disclose disabilities and to provide guidance and suggestions about appropriate support. ADMINS explores the potential of CUIs to reduce administrative burden and improve the experience of arranging support by replacing a static form with written or spoken dialogue. This article reports the results of two trials conducted during the project. A beta trial using an early version of the VA provided understanding of accessibility challenges and issues in user experience. The beta trial sample included 22 students who had already disclosed disabilities and 3 disability support advisors. After improvements to the design, a larger main trial was conducted with 134 students who disclosed their disabilities to the university using both the VA and the existing form-based process. The results show that the VA was preferred by most participants to completing the form (64.9% vs 23.9%). Qualitative and quantitative feedback from the trials also identified accessibility and user experience barriers for improving CUI design, and an understanding of benefits and preferences that can inform further development of accessible CUIs for this design space.

Conversational user interfaces (CUIs) have become an everyday technology for people the world over, as well as a booming area of research. Advances in voice synthesis and the emergence of chatbots powered by large language models (LLMs), notably ChatGPT, have pushed CUIs to the forefront of human-computer interaction (HCI) research and practice. Now that these technologies enable an elemental level of usability and user experience (UX), we must turn our attention to higher-order human factors: trust and reliance. In this workshop, we aim to bring together a multidisciplinary group of researchers and practitioners invested in the next phase of CUI design. Through keynotes, presentations, and breakout sessions, we will share our knowledge, identify cutting-edge resources, and fortify an international network of CUI scholars. In particular, we will engage with the complexity of trust and reliance as attitudes and behaviours that emerge when people interact with conversational agents.

It is well-established that more frequent social interaction is associated with higher well-being across the lifespan. The present study examines the role of frequency of interactions via different modalities on older adults' weekly well-being during the COVID-19 pandemic, where people had to adapt their communication behavior and reduce in-person contact due to precautionary measures. We use data from 98 participants (age: M = 71, SD = 5), who documented their weekly frequency of communication via four interaction modalities as well as their loneliness, positive affect, and negative affect over up to 64 weeks. Results show that participants with overall higher frequency of face-to-face, telephone, and text-based interaction than others report higher levels of positive affect and lower levels of negative affect and loneliness than others. Participants report higher levels of well-being during weeks when they report more frequent face-to-face, telephone, and text-based interaction than their individual average. Unexpectedly, participants report higher levels of negative affect during weeks with more video call interaction. Some effects of social interaction frequency on affect and loneliness are higher for face-to-face interactions versus other modalities. In addition, interaction effects at within-person level indicate that the effects of weekly telephone and text-based interaction frequency on loneliness are stronger in weeks with relatively few face-to-face interactions. Taken together, our findings suggest that social interactions via different modalities contribute to well-being, but that face-to-face interactions have the biggest effect. In addition, there is some evidence that telephone and text-based interaction may play a compensatory role. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

AimThis study aimed to predict the formation of OBL during femtosecond laser SMILE surgery by employing deep learning technology.MethodsThis was a cross-sectional, retrospective study conducted at a university hospital. Surgical videos were randomly divided into a training (3,271 patches, 73.64%), validation (704 patches, 15.85%), and internal verification set (467 patches, 10.51%). An artificial intelligence (AI) model was developed using a SENet-based residual regression deep neural network. Model performance was assessed using the mean absolute error (EMA), Pearson’s correlation coefficient (r), and determination coefficient (R2).ResultsFour distinct types of deep neural network models were established. The modified deep residual neural network prediction model with channel attention built on the PyTorch framework demonstrated the best predictive performance. The predicted OBL area values correlated well with the Photoshop-based measurements (EMA = 0.253, r = 0.831, R2 = 0.676). The ResNet (EMA = 0.259, r = 0.798, R2 = 0.631) and Vgg19 models (EMA = 0.31, r = 0.758, R2 = 0.559) both displayed satisfactory predictive performance, while the U-net model (EMA = 0.605, r = 0.331, R2 = 0.171) performed poorest.ConclusionWe used a panoramic corneal image obtained before the SMILE laser scan to create a unique deep residual neural network prediction model to predict OBL formation during SMILE surgery. This model demonstrated exceptional predictive power, suggesting its clinical applicability across a broad field.

Virtual reality (VR) technology is widely used in the field of education and cultural heritage learning. However, few studies have focused on the impact on creativity regarding the use of different interaction modalities by users in cultural heritage learning. Furthermore, it is controversial whether VR can enhance user experience and learning outcomes. The study utilized a mixed research approach combining qualitative and quantitative methods, where qualitative data was collected using open-ended interviews, simultaneously quantitative data was collected using learning outcomes, intrinsic motivation, workload, self-report of creativity and expert assessments of creativity. A mixed research approach combining qualitative and quantitative methods was used; creativity self-reports were used to collect qualitative data, and learning outcomes, intrinsic motivation, and workload data were collected for the quantitative findings.90 university students were randomly assigned to the text group (TG), the video group (VG), and the immersive virtual reality group (IVRG), with 30 students in each group. The study assessed whether creativity, user experience and learning outcomes differed when using different interaction modalities in a cultural heritage context. The results demonstrated that the text interaction modality outperformed the other two groups in learning outcomes and expert assessments of creativity performance, while the IVRG performed better in self-report of creativity and user experience. In addition, participants in TG had higher workloads and poorer user experiences compared to the other two interaction modalities. This study encourages that different interaction modalities should be selected more appropriately depending on the purpose. The study found that interaction modalities creativity can be applied to digital cultural heritage and education and informs interaction design for the HCI community.

User experience (UX) has become an important aspect in the evaluation of interactive systems. In parallel, conversational interfaces have been increasingly used in many work and everyday settings. Although there have been various methods developed to evaluate conversational interfaces, there has been a lack of methods specifically focusing on evaluating user experience. This study reviews the six main questionnaires for evaluating conversational systems in order to assess the potential suitability of these questionnaires to measure various UX dimensions. We found that (i) four questionnaires included assessment items, in varying extents, to measure hedonic, aesthetic and pragmatic dimensions of UX; (ii) two questionnaires assessed affect, and one assessed frustration dimension; and, (iii) enchantment, playfulness and motivation dimensions have not been covered sufficiently by any questionnaires. We recommend using multiple questionnaires to obtain a more complete measurement of user experience or improve the assessment of a particular UX dimension.

Deep Neural Networks (DNNs) have achieved a great success in machine learning. Among a lot of DNN structures, Deep Convolutional Neural Networks (DCNNs) are currently the main tool in the state-of-the-art variety of classification tasks like visual object recognition and handwriting and speech recognition. Despite wide perspectives, DCNNs have still some challenges to deal with. In previous work, we demonstrated the effectiveness of using some regularization techniques such as the dropout to enhance the performance of DCNNs. However, DCNNs need enough training data or even a class balance within datasets to conduct better results. To resolve this problem, some researchers have evoked different data augmentation approaches. This paper presents an extension of a later study. In this work, we conducted and compared the results of many experiments on CIFAR-10, STL-10 and SVHN using variant techniques of data augmentation combined with regularization techniques. The analysis results show that with the right use of data augmentation approaches, it is possible to achieve good results and outperform the state-of-the-art in this field.

The use of speech as an interaction modality has grown considerably through the integration of Intelligent Personal Assistants (IPAs- e.g. Siri, Google Assistant) into smartphones and voice based devices (e.g. Amazon Echo). Such engineering advances in speech processing present a unique opportunity for enabling users to interact with interface in a truly conversational way. However, we have yet to see current voice-enable interface fully becoming Conversational User Interfaces (CUIs) as afforded by the underlying speech and natural language capabilities. For example, from a conversational / dialogue perspective, there remain significant gaps in using theoretical frameworks to understand user behaviours and choices and how they may applied to specific speech interface interactions. On a design and Human-Computer Interaction level, we don't yet have the proper tools such as validated design guidelines to help us improve the usability of such interfaces. On the speech processing side, variability in speech, language, and conversation still pose problem, and error-recovery strategies often lead to degraded user experience. From a critical perspective, issues of ethics and privacy remain yet to be addressed.