Agent-based training systems can enhance people's social skills. The effective development of these systems needs a comprehensive architecture that outlines their components and relationships. Such an architecture can pinpoint improvement areas and future outlooks. This paper presents ARTES: a general architecture illustrating how components of agent-based social training systems work together. We studied existing systems and architectures for training and tutoring to design ARTES and identify its essential components and interaction characteristics. ARTES comprises two core components: the agent simulation of social situations, and educational elements to provide guided learning. We link ARTES's crucial components to four primary learning theories (behaviourism, cognitivism, social cognitive theory, and constructivism) to illustrate the role of agent simulation and tutoring elements in establishing desired learning outcomes. Furthermore, we map ARTES's components against eight architectures, 43 systems and three tools to indicate the components' relevance, completeness, generalisation, and deployment potential across contexts. In addition to ARTES, the paper also contributes by identifying future improvements and research directions, such as the agent's thinking, tutoring methods, knowledge transfer, and ethical implications. We believe ARTES can help bridge the gap between virtual human simulations and impactful educational learning, offering training system developers desirable features like understandability and adaptability.@en

Goal-setting is commonly used in behavior change applications for physical activity. However, for goals to be effective, they need to be tailored to a user’s situation (e.g., motivation, progress). One way to obtain such goals is a collaborative process in which a healthcare professional and client set a goal together, thus making use of the professional’s expertise and the client’s knowledge about their own situation. As healthcare professionals are not always available, we created a dialog with the virtual coach Steph to collaboratively set daily step goals. Since judgments in human decision-making processes are adjusted based on the starting point or anchor, the first step goal proposal Steph makes is likely to influence the user’s final goal and self-efficacy. Situational factors impacting physical activity (e.g., motivation, self-efficacy, available time) or how users process information (e.g., mood) may determine which initial proposals are most effective in getting users to reach their underlying previous activity-based recommended step goals. Using data from 117 people interacting with Steph for up to five days, we designed a reinforcement learning algorithm that considers users’ current and future situations when choosing an initial step goal proposal. Our simulations show that initial step goal proposals matter: choosing optimal ones based on this algorithm could make it more likely that people move to a situation with high motivation, high self-efficacy, and a favorable daily context. Then, they are more likely to achieve, but also to overachieve, their underlying recommended step goals. Our dataset is publicly available.@en

In the published article, there was an error in Table 5. For each second construct/dimension, the means are swapped between Chinese and English data, which is caused by an error in the underlying R script. Consequently, the plus and minus signs for the delta and CI values are also wrong. The corrected Table 5 and its caption appear below. Construct/dimension rating difference between mixed-international English-speaking and Chinese mother-tongue groups. Δ Score are pairwise differences between Chinese and mother-tongue cultural background and mixed-international cultural background taken from the posterior distribution. M, mean; SD, standard deviation; CI, credible interval. The authors apologize for this error and state that this does not change the scientific conclusions of the article in any way. The original article has been updated.

@en

To increase the effectiveness of behavior change applications, a large variety of algorithms has been developed to adapt what the applications offer, when, how, and with whom. Given the multitude of challenges related to the concept of algorithmic behavior change support, its development, evaluation, and impact on behavior change, this workshop aims to strengthen the community of people with diverse backgrounds (e.g., computer science, psychology, human-computer interaction) and roles in behavior change support (e.g., researcher, designer, practitioner). Combining keynotes of leading researchers with sessions in which individual workshop participants present their work and discuss problems with the audience, the workshop encouraged a lively exchange of ideas that benefits current and future research on algorithmic behavior change support.@en

Enabling the widespread utilization of the Artificial-Social-Agent (ASA) Questionnaire, a research instrument to comprehensively assess diverse ASA qualities while ensuring comparability, necessitates translations beyond the original English source language questionnaire. We thus present Dutch and German translations of the long and short versions of the ASA Questionnaire and describe the translation challenges we encountered. Summative assessments with 240 English-Dutch and 240 English-German bilingual participants show, on average, excellent correlations (Dutch ICC M = 0.82, SD = 0.07, range [0.58, 0.93]; German ICC M = 0.81, SD = 0.09, range [0.58, 0.94]) with the original long version on the construct and dimension level. Results for the short version show, on average, good correlations (Dutch ICC M = 0.65, SD = 0.12, range [0.39, 0.82]; German ICC M = 0.67, SD = 0.14, range [0.30, 0.91]). We hope these validated translations allow the Dutch and German-speaking populations to evaluate ASAs in their own language.@en

Social interactions form an essential aspect of people’s life, however, it is quite challenging for individuals to handle a wide range of social situations. Therefore, a variety of training systems have been developed to improve their skills. This literature review seeks to give an overview of the state of the art of technology-supported systems for social skills training. The studies eligible for inclusion described a technology-supported system with the purpose of training social skills and included an experimental or observational study to evaluate the efficacy of the system. 225 studies (224 publications) with 216 systems were identified, characterized, and analyzed in this literature review. Using the taxonomy as put forward in this study, the analysis shows that the majority of these systems were screen-based applications, with virtual reality technology being the most frequently observed. The systems most often targeted communication skills that focus on transferring information to produce greater understanding, i.e. mending general communication impairments in children with autism. In terms of functions, support for learning-by-doing was the most observed function, while focusing on job interviews provided the largest number of functions. Finally, the studies reported overwhelmingly positively regarding the systems’ impact, including 76 studies with a randomized controlled trial design. Still, most studies only used a quasi-experimental design based on self-report measures. We anticipate the proposed taxonomy to be a starting point for researchers to position their work and that the review will help them with gaining inspiration for the design and evaluation of social skills training systems.@en

This document is an encore abstract of the paper “Persuading to Prepare for Quitting Smoking with a Virtual Coach: Using States and User Characteristics to Predict Behavior” presented at AAMAS 2023.@en

Despite their prevalence in eHealth applications for behavior change, persuasive messages tend to have small effects on behavior. Conditions or states (e.g., confidence, knowledge, motivation) and characteristics (e.g., gender, age, personality) of persuadees are two promising components for more effective algorithms for choosing persuasive messages. However, it is not yet sufficiently clear how well considering these components allows one to predict behavior after persuasive attempts, especially in the long run. Since collecting data for many algorithm components is costly and places a burden on users, a better understanding of the impact of individual components in practice is welcome. This can help to make an informed decision on which components to use. We thus conducted a longitudinal study in which a virtual coach persuaded 671 daily smokers to do preparatory activities for quitting smoking and becoming more physically active, such as envisioning one’s desired future self. Based on the collected data, we designed a Reinforcement Learning (RL)-approach that considers current and future states to maximize the effort people spend on their activities. Using this RL-approach, we found, based on leave-one-out cross-validation, that considering states helps to predict both behavior and future states. User characteristics and especially involvement in the activities, on the other hand, only help to predict behavior if used in combination with states rather than alone. We see these results as supporting the use of states and involvement in persuasion algorithms. Our dataset is available online.@en

Background: People with diabetes mellitus not only have to deal with physical health problems, but also with the psycho-social challenges their chronic disease brings. Currently, technological tools that support the psycho-social context of a patient have received little attention.

Objective: The objective of this work is to determine the feasibility and preliminary efficacy of an automated conversational agent to deliver, to people with diabetes, personalised psycho-education on dealing with (psycho-)social distress related to their chronic illness.

Methods: In a double-blinded between-subject study, 156 crowd-workers with diabetes received a social help program intervention in three sessions over three weeks. They were randomly assigned to receive support from either an interactive conversational support agent (n=79) or a self-help text from the book “Diabetes burnout” as a control condition (n=77). Participants completed the Diabetes Distress Scale (DDS) before and after the intervention, and after the intervention, the Client Satisfaction Questionnaire (CSQ-8), Feeling of Being Heard (FBH), and System Usability Scale (SUS).

Results: Results indicate that people using the conversational agent have a larger reduction in diabetes distress (M=−0.305, SD=0.865) than the control group (M=0.002, SD=0.743) and this difference is statistically significant (t(154)=2.377, p=0.019). A hypothesised mediation effect of “attitude to the social help program” was not observed.

Conclusions: An automated conversational agent can deliver personalised psycho-education on dealing with (psycho-)social distress to people with diabetes and reduce diabetes distress more than a self-help book.

Ethics, Study Registration and Open Science: This study has been preregistered with the Open Science Foundation (osf.io/yb6vg) and has been accepted by the Human Research Ethics Committee - Delft University of Technology under application number 1130. The data and analysis script are available: https://surfdrive.surf.nl/files/index.php/s/4xSEHCrAu0HsJ4P.@en

Counsellors at the child helpline offer a confidential environment for children to be heard and empowered. However, training counsellors on handling children’s conversations in text-based chat can be costly and time-consuming. This paper introduces Lilobot, a conversational agent designed for training counsellors of child helplines. The agent’s dialogue is built on the Belief-Desire-Intention (BDI) model, which, in this case, simulates a child victim of school bullying in a text based interaction. Trainees engage with Lilobot in a role-play format, taking on the counsellor’s role. This interactive system helps trainees learn the Five Phase Model, a conversation protocol child’s helplines use. The system also has a trainer interface, where a trainer can oversee and control Lilobot’s interactions, and see a suggested optimal conversational path. The system was built with three main components - a natural language processing model (using Rasa) and the BDI reasoning model and optimal path generation (using Java Spring).@en

Tailored motivational messages are helpful to motivate people in eHealth applications for increasing physical activity, but it is not sufficiently clear how such messages can be effectively generated in advance. We, therefore, put forward a theory-driven approach to generating tailored motivational messages for eHealth applications for behavior change, and we examine its feasibility by assessing how motivating the resulting messages are perceived. For this, we designed motivational messages with a specific structure that was based on an adaptation of an existing ontology for tailoring motivational messages in the context of physical activity. To obtain tailored messages, experts in health psychology and coaching successfully wrote messages with this structure for personas in scenarios that differed with regard to the persona’s mood, self-efficacy, and progress. Based on an experiment in which 60 participants each rated the perceived motivational impact of six generic and six tailored messages based on scenarios, we found credible support for our hypothesis that messages tailored to mood, self-efficacy, and progress are perceived as more motivating. A thematic analysis of people’s free-text responses about what they found motivating and demotivating about motivational messages further supports the use of tailored messages, as well as messages that are encouraging and empathetic, give feedback about people’s progress, and mention the benefits of physical activity. To aid future work on motivational messages, we make our motivational messages and corresponding scenarios publicly available.@en

Background: While research on Virtual Reality’s potential for education continues to advance, research on its support for Collaborative Learning is small in scope. With remote collaboration and distance learning becoming increasingly relevant for education (especially since the COVID-19 pandemic), an understanding of Virtual Reality’s potential for Collaborative Learning is of importance. To establish how this immersive technology can support and enhance collaboration between learners, this systematic literature review analyses scientific research on Virtual Reality for Collaborative Learning with the intention to identify 1) skills and competences trained, 2) domains and disciplines addressed, 3) systems used and 4) empirical knowledge established.Method: Two scientific databases—Scopus and Web of Science—were used for this review. Following the PRISMA method, a total of 139 articles were analyzed. Reliability of this selection process was assessed using five additional coders. A taxonomy was used to classify these articles. Another coder was used to assess the reliability of the primary coder before this taxonomy was applied to the selected articlesResults: Based on the literature reviewed, skills and competences developed are divided into five categories. Educational fields and domains seem interested in Virtual Reality for Collaborative Learning because of a need for innovation, communities and remote socialization and collaboration between learners. Systems primarily use monitor-based Virtual Reality and mouse-and-keyboard controls. A general optimism is visible regarding the use of Virtual Reality to support and enhance Collaborative LearningConclusion: Five distinct affordances of Virtual Reality for Collaborative Learning are identified: it 1) is an efficient tool to engage and motivate learners, 2) supports distance learning and remote collaboration, 3) provides multi- and interdisciplinary spaces for both learning and collaborating, 4) helps develop social skills and 5) suits Collaborative Learning-related paradigms and approaches. Overall, the reviewed literature suggests Virtual Reality to be an effective tool for the support and enhancement of Collaborative Learning, though further research is necessary to establish pedagogies.@en

Goal-setting is often used in eHealth applications for behavior change as it motivates and helps to stay focused on a desired outcome. However, for goals to be effective, they need to meet criteria such as being specific, measurable, attainable, relevant and time-bound (SMART). Moreover, people need to be confident to reach their goal. We thus created a goal-setting dialog in which the virtual coach Jody guided people in setting SMART goals. Thereby, Jody provided personalized vicarious experiences by showing examples from other people who reached a goal to increase people’s confidence. These experiences were personalized, as it is helpful to observe a relatable other succeed. Data from an online study with a between-subjects with pre-post measurement design (n=39 participants) provide credible support that personalized experiences are seen as more motivating than generic ones. Motivational factors for participants included information about the goal, path to the goal, and the person who accomplished a goal, as well as the mere fact that a goal was reached. Participants also had a positive attitude toward Jody. We see these results as an indication that people are positive toward using a goal-setting dialog with a virtual coach in eHealth applications for behavior change. Moreover, contrary to hypothesized, our observed data give credible support that participants’ self-efficacy was lower after the dialog than before. These results warrant further research on how such dialogs affect self-efficacy, especially whether these lower post-measurements of self-efficacy are associated with people’s more realistic assessment of their abilities.@en

Virtual coaches have the potential to address the low adherence common to eHealth applications for behavior change by, for example, providing motivational support. However, given the multitude of factors affecting users’ attitudes toward virtual coaches, more insights are needed on how such virtual coaches can be designed to affect these attitudes in a specific use context positively. Especially valuable are insights that are based on users interacting with such a virtual coach for longer. We thus conducted a study in which more than 500 smokers interacted with the text-based virtual coach Sam in five sessions. In each session, Sam assigned smokers a new preparatory activity for quitting smoking and provided motivational support for doing the activity. Based on a mixed-methods analysis of users’ willingness to continue working and their relationship with Sam, we obtained eight themes for users’ attitudes toward Sam. These themes relate to whether Sam is seen as human or artificial, specific characteristics of Sam (e.g., caring character), the interaction with Sam, and the relationship with Sam. We used these themes to formulate literature-based recommendations to guide designers of virtual coaches for behavior change. For example, letting the virtual coach get to know users and disclose more information about itself may improve its relationship with users.@en

Background: Globally, suicide is among the leading causes of death, with men being more at risk to die from suicide than women. Research suggests that people with suicidal ideation often struggle to find adequate help. Every month, around 4000 people fill in the anonymous self-test for suicidal thoughts on the website of the Dutch suicide prevention helpline. This self-test includes the Suicidal Ideation Attributes Scale (SIDAS), which educates users about the severity of their suicidal thoughts. The vast majority (70%) of people who complete the self-test score higher than the cutoff point (≥21) for severe suicidal thoughts. Unfortunately, despite this, less than 10% of test-takers navigate to the web page about contacting the helpline. Objective: This protocol presents the design of a web-based randomized controlled trial that aims to reduce barriers to contacting the suicide prevention helpline. The aim of this study is 2-fold: (1) to measure the effectiveness of a brief barrier reduction intervention (BRI) provided in the self-test motivating people with severe suicidal thoughts to contact the Dutch suicide prevention helpline and (2) to specifically evaluate the effectiveness of the BRI in increasing service use by high-risk groups for suicide such as men and middle-aged people. Methods: People visiting the self-test for suicidal thoughts on the website of the suicide prevention helpline will be asked to participate in a study to improve the self-test. Individuals with severe suicidal thoughts and little motivation to contact the helpline will be randomly allocated either to a brief BRI, in which they will receive a short tailored message based on their self-reported barrier to the helpline (n=388) or care as usual (general advisory text, n=388). The primary outcome measure is the use of a direct link to contact the helpline after receiving the intervention or control condition. Secondary outcomes are the self-reported likelihood of contacting the helpline (on a 5-point scale) and satisfaction with the self-test. In the BRI, participants receive tailored information to address underlying concerns and misconceptions of barriers to the helpline. A pilot study was conducted among current test-takers to identify these specific barriers. Results: The pilot study (N=1083) revealed multiple barriers to contacting the helpline. The most prominent were the belief that a conversation with a counselor would not be effective, fear of the conversation itself, and emotional concerns about talking about suicidal thoughts. Conclusions: Our study will provide insight into the effectiveness of a brief BRI designed to increase the use of a suicide prevention helpline provided in a self-test on suicidal thoughts. If successful, this intervention has the potential to be a low-cost, easily scalable, and feasible method to increase service use for helplines across the world.

@en

The Artificial-Social-Agent (ASA) questionnaire is an instrument for evaluating human-ASA interaction. It consists of 19 constructs and related dimensions measured by either 24 questionnaire items (short version) or 90 questionnaire items (long version). The questionnaire was built and validated by a research community effort to make evaluation results more comparable between agents and findings more generalizable. The current questionnaire is in English, which limits its use to only a population with an adequate command of the English language. Translating the questionnaire into more languages allows for the inclusion of other populations and the possibility of comparing them. Therefore, this paper presents a Mandarin Chinese translation of the questionnaire. After three construction cycles that included forward and backward translation, we gave both the final version of the translated and original English questionnaire to 242 bilingual crowd-workers to evaluate 14 ASAs. Results show on average a good level of correlation on the construct/dimension level (ICC M = 0.79, SD = 0.09, range [0.61, 0.95]) and on the item level (ICC M = 0.62, SD = 0.14, range [0.19, 0.92]) between the two languages for the long version, and for the short version (ICC M = 0.66, SD = 0.12, range [0.41, 0.92]). The analysis also established correction values for converting questionnaire item scores between Chinese and English questionnaires. Moreover, we also found systematic differences in English questionnaire scores between the bilingual sample and a previously collected mixed-international English-speaking sample. We hope this and the Chinese questionnaire translation will motivate researchers to study human-ASA interaction among a Chinese literate population and to study cultural similarities and differences in this area.@en

We present the ASA Questionnaire, an instrument for evaluating human interaction with an artificial social agent (ASA), resulting from multi-year efforts involving more than 100 Intelligent Virtual Agent (IVA) researchers worldwide. It has 19 measurement constructs constituted by 90 items, which capture more than 80% of the constructs identified in empirical studies published in the IVA conference 2013 - 2018. This paper reports on construct validity analysis, specifically convergent and discriminant validity of initial 131 instrument items that involved 532 crowd-workers who were asked to rate human interaction with 14 different ASAs. The analysis included several factor analysis models and resulted in the selection of 90 items for inclusion in the long version of the ASA questionnaire. In addition, a representative item of each construct or dimension was selected to create a 24-item short version of the ASA questionnaire. Whereas the long version is suitable for a comprehensive evaluation of human-ASA interaction, the short version allows quick analysis and description of the interaction with the ASA. To support reporting ASA questionnaire results, we also put forward an ASA chart. The chart provides a quick overview of the agent profile.

@en

Virtual patients (VPs) offer an affordable and feasible method to train individuals when compared to patient-actors. They can provide training on communication skills, such as motivational interviewing and conflict resolution, and often facilitate a change in patients’ thinking and emotional states. However, few studies have focused on VPs with cognitive and emotional states and internal schema or rules that govern them. In this research, we aim to design and empirically study a VP training system with a mental model to enrich the interactions and training. With such a model, the learning environment has the potential to generate valuable feedback and guidance for the learner based on the states of the VP. We started by examining systems aimed at training individuals with a virtual agent that simulates a person in a social situation (e.g., a virtual customer to train salespersons). We developed an architecture for these systems, defined the current interaction approaches, and will discuss the main aspects of the training system architecture and various technological approaches. Here, we consider the potential solutions proposed in adjacent virtual-agent domains. Based on our findings, we will model the VP’s reasoning to improve trainees’ communication skills and investigate how helpful feedback and guidance could be generated from these mental models at run time. The main contribution of our work will be the build of an empirically grounded virtual patient with a mental model as a training system. @en

E-mental health for depression is increasingly used in clinical practice, but patient adherence suffers as therapist involvement decreases. One reason may be the low responsiveness of existing programs: especially autonomous systems are lacking in their input interpretation and feedback-giving capabilities. Here, we explore (a) to what extent a more socially intelligent and, therefore, technologically advanced solution, namely a conversational agent, is a feasible means of collecting thought record data in dialog, (b) what people write about in their thought records, (c) whether providing content-based feedback increases motivation for thought recording, a core technique of cognitive therapy that helps patients gain an understanding of how their thoughts cause their feelings. Using the crowd-sourcing platform Prolific, 308 participants with subclinical depression symptoms were recruited and split into three conditions of varying feedback richness using the minimization method of randomization. They completed two thought recording sessions with the conversational agent: one practice session with scenarios and one open session using situations from their own lives. All participants were able to complete thought records with the agent such that the thoughts could be interpreted by the machine learning algorithm, rendering the completion of thought records with the agent feasible. Participants chose interpersonal situations nearly three times as often as achievement-related situations in the open chat session. The three most common underlying schemas were the Attachment, Competence, and Global Self-evaluation schemas. No support was found for a motivational effect of providing richer feedback. In addition to our findings, we publish the dataset of thought records for interested researchers and developers.
@en