For personal assistive technologies to effectively support users, they need a user model that records information about the user, such as their goals, values, and context. Knowledge-based techniques can model the relationships between these concepts, enabling the support agent to act in accordance with the user's values. However, user models require updating over time to accommodate changes and continuously align with what the user deems important. In our work, we propose and investigate the use of human-agent alignment dialogues for establishing whether user model updates are needed and acquiring the necessary information for these updates. In this paper, we perform an exploratory qualitative focus group study in which we investigate participants' opinions about written examples of alignment dialogues, as a foundation for their design. Transcripts were analyzed using thematic analysis. A main theme that emerged concerns the potential impact of agent utterances on the user's feelings about themselves and about the agent.

@en

Hybrid Intelligence (HI) is the combination of human and machine intelligence, expanding human intellect instead of replacing it. Information in HI scenarios is often inconsistent, e.g. due to shifting preferences, user's motivation or conflicts arising from merged data. As it provides an intuitive mechanism for reasoning with conflicting information, with natural explanations that are understandable to humans, our hypothesis is that Dung's Abstract Argumentation (AA) is a suitable formalism for such hybrid scenarios. This paper investigates the capabilities of Argumentation in representing and reasoning in the presence of inconsistency, and its potential for intuitive explainability to link between artificial and human actors. To this end, we conduct a survey among a number of research projects of the Hybrid Intelligence Centre. Within these projects we analyse the applicability of argumentation with respect to various inconsistency types stemming, for instance, from commonsense reasoning, decision making, and negotiation. The results show that 14 out of the 21 projects have to deal with inconsistent information. In half of those scenarios, the knowledge models come with natural preference relations over the information. We show that Argumentation is a suitable framework to model the specific knowledge in 10 out of 14 projects, thus indicating the potential of Abstract Argumentation for transparently dealing with inconsistencies in Hybrid Intelligence systems.@en

Artificial agents that support people in their daily activities (e.g., virtual coaches and personal assistants) are increasingly prevalent. Since many daily activities are social in nature, support agents should understand a user's social situation to offer comprehensive support. However, there are no systematic approaches for developing support agents that are social situation aware. We identify key requirements for a support agent to be social situation aware and propose steps to realize those requirements. These steps are presented through a conceptual architecture centered on two key ideas: 1) conceptualizing social situation awareness as an instantiation of "general" situation awareness, and 2) using situation taxonomies for such instantiation. This enables support agents to represent a user's social situation, comprehend its meaning, and assess its impact on the user's behavior. We discuss empirical results supporting the effectiveness of the proposed approach and illustrate how the architecture can be used in support agents through two use cases.

@en

Effective psychological interventions for anxiety disorders often include exposure to fearful situations. However, individuals with low self-efficacy may find such exposure too overwhelming. We created a vicarious experience in virtual reality, which enables observation of one’s experience from a first person perspective without actual performance and which might increase self-efficacy. With similarities to both traditional vicarious experiences and direct experiences, the level of self-identification with the experience was hypothesized to affect self-efficacy and its relationship with direct experiences. To test this, vicarious experiences with two distinct levels of self-identification were compared in a between-subjects experiment ((Formula presented.)). After being exposed to a vicarious experience of giving lectures on elementary arithmetic in front of a virtual audience with either a high or low level of self-identification with the public speaker, participants from both conditions actively gave another lecture. The results revealed that self-identification affected people’s self-efficacy after vicarious experience. They further revealed that self-identification is a moderator of (1) the correlation between perceived performance and self-efficacy, (2) the correlation between self-efficacy measured after the vicarious and the follow-up direct experience; and (3) the correlation between the sense of presence reported in the vicarious and in the follow-up direct experience. We anticipate that the first-person-perspective experiences with high-level of self-identification have the potential to be beneficial for training where changing people’s self-efficacy is desirable.

@en

Personal assistant agents have been developed to help people in their daily lives with tasks such as agenda management. In order to provide better support, they should not only model the user’s internal aspects, but also their social situation. Current research on social context tackles this by modelling the social aspects of a situation from an objective perspective. In our approach, we model these social aspects of the situation from the user’s subjective perspective. We do so by using concepts from social science, and in turn apply machine learning techniques to predict the priority that the user would assign to these situations. Furthermore, we show that using these techniques allows agents to determine which features influenced these predictions. Results based on a crowd-sourcing user study suggest that our proposed model would enable personal assistant agents to differentiate between situations with high and low priority. We believe this to be a first step towards agents that better understand the user’s social situation, and adapt their support accordingly.

@en

Support agents are investigated more and more as a way of assisting people in carrying out daily tasks. Support agents should be flexible in adapting their support to what their user needs. Research suggests that the situation someone is in affects their behaviour, however its effect has not been incorporated in the decision making of support agents. Modelling the characteristics of situations explicitly and studying their effect on internal perceptions of the user, such as their personal values, would enable support agents to provide more personalized support. We propose a method which groups situations according to their psychological characteristics, and in turn determines which personal values of the user would be promoted or demoted in each group of situations. To do this, we conduct a user study to gather data from participants about situations that they encounter in their daily lives. Results show that the created groups of situations significantly promote or demote certain personal values. This approach can allow support agents to help the user in a way which is in line with their personal values. @en

We define hybrid intelligence (HI) as the combination of human and machine intelligence, augmenting human intellect and capabilities instead of replacing them and achieving goals that were unreachable by either humans or machines. HI is an important new research focus for artificial intelligence, and we set a research agenda for HI by formulating four challenges.@en

Behaviour support agents need to be aware of the social environment of the user in order to be able to provide comprehensive support. However, this is a feature that is currently lacking in existing systems. To tackle it, first of all we explore literature from social sciences in order to find which elements of the social environment need to be represented. We structure this knowledge as a two-level ontology that models social situations. We formalize the elements that are needed to model social situations, which consist of different types of meetings between two people. We conduct an experiment to evaluate the lower level of the ontology using feedback from the subjects, and to test whether we can use the data to reason about the priority of different situations. Subjects found our proposed features of social relationships to be understandable and representative. Furthermore, we show these features can be combined in a decision tree to predict the priority of social situations.

@en

Agents in teamwork may be highly interdependent on each other, the awareness of interdependence relationships is an important requirement for designing and consequently implementing a multi-agent system. In this work, we propose a formal graphical and domain-independent language that can facilitate the identification of comprehensive interdependences among the agents in teamwork. Moreover, a formal semantics is also introduced to precisely express and explain the properties of a graphical structure. The novel feature of the graphical language is that it complements the Interdependence Analysis Color Scheme in a way that explicitly models negative influences and, in addition, provides a visual-communication aid for developers. To demonstrate the applicability and sufficiency of the graphical language in a variety of domains, our case studies include a multi-robot scenario and a human-robot scenario.

@en

HRI researchers have explored how people behave toward technology agents, advancing the concept that people can attain 'closeness' with technology itself in addition to a living social partner. Yet the topic of closeness with robots has not been fully explored or organized into a discrete area of study. This seems particularly important to the design of robots that are expressive, to the implementation of technologies that use new social signal processing or reciprocal social touch, and to the study of how people respond to robots. This half-day workshop is a forum to discuss the future of 'closeness' with robots, conversational agents, autonomous vehicles, Internet of Things devices and other technologies that act as social partners - designs, applications, responses and societal concerns.

@en

Behaviour support technology assists people in organising their daily activities and changing their behaviour. A fundamental notion underlying such supportive technology is that of compliance with behavioural norms: do people indeed perform the desired behaviour? Existing technology employs a rigid implementation of compliance: a norm is either satisfied or not. In practice however, behaviour change norms are less strict: E.g., is a new norm to do sports at least three times a week complied with if it is occasionally only done twice a week? To address this, in this paper we formally specify probabilistic norms through a variant of feature diagrams, enabling a hierarchical decomposition of the desired behaviour and its execution frequencies. Further, we define a new notion of probabilistic norm compliance using a formal hypothesis testing framework. We show that probabilistic norm compliance can be used in a real-world setting by implementing and evaluating our semantics with respect to an existing daily behaviour dataset.

@en

Personal technology such as electronic partners (e-partners) play an increasing role in our daily lives, and can make an important difference by supporting us in various ways. However, when they offer this support, it is important that they do so with an understanding of our choices and what is important to us. To allow an e-partner to flexibly do this, we propose a formal framework to automatically derive norms which describe how to perform a certain behavior. These norms are directly derived from the user's actions, values and the context they are in. In this way, the e-partner can take into account the user's values and offer more flexible personalized support.

@en

Behaviour support technology is aimed at assisting people in organizing their Activities of Daily Living (ADLs). Numerous frameworks have been developed for activity recognition and for generating specific types of support actions, such as reminders. The main goal of our research is to develop a generic formal framework for representing and reasoning about ADLs and their temporal relations. This framework should facilitate modelling and reasoning about 1) durative activities, 2) relations between higher-level activities and subactivities, 3) activity instances, and 4) activity duration. In this paper we present a temporal logic as an extension of the logic TPTL for specification of real-time systems. Our logic TPTL_bih is defined over Behaviour Identification Hierarchies (BIHs) for representing ADL structure and typical activity duration. To model execution of ADLs, states of the temporal traces in TPTL_bih comprise information about the start, stop and current execution of activities. We provide a number of constraints on these traces that we stipulate are desired for the accurate representation of ADL execution, and investigate corresponding validities in the logic. To evaluate the expressivity of the logic, we give a formal definition for the notion of Coherence for (complex) activities, by which we mean that an activity is done without interruption and in a timely fashion. We show that the definition is satisfiable in our framework. In this way the logic forms the basis for a generic monitoring and reasoning framework for ADLs.

@en

Social commitments (SCs) provide a flexible, norm-based, governance structure for sharing and receiving data. However, users of data sharing applications can subscribe to multiple SCs, possibly producing opposing sharing and receiving requirements. We propose resolving such conflicts automatically through a conflict resolution model based on relevant user values such as privacy and safety. The model predicts a user’s preferred resolution by choosing the commitment that best supports the user’s values. We show through an empirical user study (n = 396) that values, as well as recency and norm type, significantly improve a system’s ability to predict user preference in location sharing conflicts.@en

Mobile location-sharing technology is increasingly being used by parents to locate their children. Research shows that these technologies may pose risks to important user values such as privacy and responsibility, while they aim to promote others such as family security. As a solution, we proposed the use of Social Commitment (SC) models for governing the sharing and receiving of data. A social commitment represents an agreement between two people about which data should (not) be shared and received in which situation. We hypothesize that the use of SCs in mobile location sharing applications provides improved support for user values since it allows for a more flexible, context-aware location sharing. In this paper, we present a user study to test this hypothesis. The study focuses on primary school children ([Formula presented]) as the main target group, who's values may be demoted through the use of location-sharing technology. Children were provided with two versions of a mobile location sharing app: one with basic check-in functionality –the basic app –and one augmented with an SC model, which we call a Socially Adaptive Electronic Partner (SAEP). Our findings suggest, among other things that the SAEP would provide improved support for children's values compared to the basic app.

@en

An institution typically comprises constitutive rules, which give shape and meaning to social interactions and regulative rules, which prescribe agent behaviour in the society. Regulative rules guide social interaction, in particular when they are coupled with reward and punishment regulations that are enforced for (non-)compliance. Institution examples include legislation and contracts. Formal institutional reasoning frameworks automate ascribing social meaning to agent interaction and determining whether those actions have social meanings that comprise (non-)compliant behaviour. Yet, institutions do not just govern societies. Rather, in what is called multi-level governance, institutional designs at lower governance levels (e.g., national legislation at the national level) are governed by higher level institutions (e.g., directives, human rights charters and supranational agreements). When an institution design is found to be non-compliant, punishments can be issued by annulling the legislation or imposing fines on the responsible designers (i.e., government). In order to enforce multi-level governance, higher governance levels (e.g., courts applying human rights) must check lower level institution designs (e.g., national legislation) for compliance; in order to avoid punishment, lower governance levels (e.g., national governments) must check their institution designs are compliant with higher-level institutions before enactment. However, checking non-compliance of institution designs in multi-level governance is non-trivial. In particular, because institutions in multi-level governance operate at different levels of abstraction. Lower level institutions govern with concrete regulations whilst higher level institutions typically comprise increasingly vague and abstract regulations. To address this issue, in this paper we propose a formal framework with a novel semantics that defines compliance between concrete lower level institutions and abstract higher level institutions. The formal framework is complemented by a sound and complete computational framework that automates compliance checking, which we apply to a real-world case study.@en

Expressive virtual audiences are used in scientific research, psychotherapy, and training. To create an expressive virtual audience, developers need to know how specific audience behaviors are associated with certain characteristics of an audience, such as attitude, and how well people can recognize these characteristics. To examine this, four studies were conducted on a virtual audience and its behavioral models: (I) a perception study of a virtual audience showed that people (n = 24) could perceive changes in some of the mood, personality, and attitude parameters of the virtual audience; (II) a design experiment whereby individuals (n = 24) constructed 23 different audience scenarios indicated that the understanding of audience styles was consistent across individuals, and the clustering of similar settings of the virtual audience parameters revealed five distinct generic audience styles; (III) a perception validation study of these five audience styles showed that people (n = 100) could differentiate between some of the styles, and the audience's attentiveness was the most dominating audience characteristic that people perceived; (IV) the examination of the behavioral model of the virtual audience identified several typical audience behaviors for each style. We anticipate that future developers can use these findings to create distinct virtual audiences with recognizable behaviors.@en