Visitors of public security facilities (i.e. detention centres, courthouses, penitentiaries and customs) often experience anxiousness and low mood during their time in the waiting room, while staff have limited capacity to attend to visitors. To address this, we explore the design of a social robot that engages visitors through socially appropriate interactions intended to support a positive mood, while simultaneously providing staff with remote situation awareness. Simultaneously, social robots have been deployed as reception agents and mood-intervention tools in public settings (e.g., healthcare[1] and retail[2]). Usually, social robots in public buildings handle welcoming and giving information; this thesis will test what happens when a reception robot also provides visitors a mood-intervention activity. A social robot could therefore provide a friendlier atmosphere in the emotionally fraught visitor area of public security constrained facilities. Here we identify three complementary robot roles – receptionist, storyteller, and playmate – combined with a monitoring function for staff. Receptionist and storyteller roles were lab-evaluated using a local LLM powered social robot as a mood-intervention agent in two iterations with a Before-After Control-Impact (BACI) study. In the lab, the design was implemented with a Softbank Pepper-robot, a pre-post study indicated that interaction effectiveness depends on turn-taking level: high turn-taking activities produce larger immediate mood gains than low turn-taking storytelling. The playmate role was field-tested in a security-constrained public service waiting room. Beyond visitor surveys, semi-structured interviews explored staff acceptance. In the real-world evaluation, the design was implemented with an Enchanted Tools Miroki robot. The results indicate that visitors particularly value the playmate role, the guessing game agent was well accepted: 89% of visitors self-reported that the robot improved their visit(t(9) = 4.0, p < .005). Staff interviews underline acceptance rests on sociotechnical framing, perceived operational usefulness and minimizing of operational overhead. The results advance understanding of social robots in high-consequence domains and other affectively complex (anxiety-laden) public places, suggesting that rapid, reciprocal interaction designs improve mood-intervention outcomes for security constrained facility visitors. Overall, the study findings inform the iterative design of social robots for sensitive, high-security public environments. Our findings suggest further elaborating the waiting room robot with a (1) time-aware conversation system to offer activities suited to visitors’ time-frames and (2) multilingual functionality for visitors without Dutch proficiency. Additionally, the other areas of the security constrained facility provide possibilities where robots may help reduce staff burden; (3) as a multilingual interpreter and as a (4) supporting buddy to reflect on accidents during visit to the facility.
... Read more

Deep-learning (DL) models could greatly advance the automation of fact-checking, yet have not widely been adopted by the public because of their hard-to-explain nature. Although various techniques have been proposed to use local explanations for the behaviour of DL models, little attention has been paid to global explanations.
In response, we investigate whether a frequent sequence mining (FSM) tool finds sequence patterns, that act as shortcuts, to a state-of-the-art model in the context of fact-checking. By studying the connections between a model’s input and output, association rules (ARs) can be used as a global explanation for the interpretation of the model. The shortcuts were evaluated using a heuristic-based minimum support value, the strength of each rule was determined using confidence, and the support value indicates the global coverage of rules. Shortcuts help to form an interpretation for creating counterfactual prompts, which can be used as a risk assessment tool for DL models. Other applications for rule-based global explanations are left for future work ... Read more