Perceptions of the Helpfulness of Unexpected Agent Assistance

Candon, Kate; Hsu, Zoe; Kim, Yoony; Chen, Jesse; Tsoi, Nathan; Vázquez, Marynel

doi:10.1145/3527188.3561915

Cited by 5 publications

(4 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While in some cases these signals are explicit feedback (e.g., to interrupt an agent [39]), our work considers them implicit feedback because we do not prime humans to react in specific ways to a robot. As such, our work is closer to [2], [37], [40]- [42], which used nonverbal signals to identify critical states during robot operation, detect robot errors, and adjust robot behavior. Other types of feedback signals, such as those from brain-computer interfaces, have been used in HRI [43]- [45]; however, they are impractical for navigation tasks.…”

Section: Related Workmentioning

confidence: 97%

“…Learning from implicit feedback is not without challenges, however, as it can be difficult to interpret [2], [3]. For example, this can happen due to inter-person variability in facial expressions [36] or similar signals being produced for different reasons [37].…”

Section: Related Workmentioning

confidence: 99%

“…Also, through this question, we study the impact of two types of observations in the prediction task: observations that describe fine-grained facial expressions for a robot user; and other observations about the user, the robot and their environment. As mentioned earlier, observations of fine-grained expressions have gained popularity in recent work to infer human perceptions of an agent's behavior [2], [4], [9], [37]. Other observations (e.g.…”

Section: Problem Statement and Research Questionsmentioning

confidence: 99%

“…These features are the spatial behavior features, the robot's goal location, the occupied space near the robot, and the gaze direction of the participant -the last of which could be approximated using an estimate of the person's head orientation [73]. Because prior work suggests that it is easier to make sense of implicit human feedback in context [37], the plot was always centered on the robot, making its surroundings always visible as in Fig. 2 (left).…”

Section: A How Well Can Human Observers Predict a User's Impression O...mentioning

confidence: 99%

See 3 more Smart Citations

Recent Advances in Mobile Robot Localization in Complex Scenarios

Zhang

Liu

et al. 2023

Lecture Notes in Electrical Engineering

View full text Add to dashboard Cite

Human impressions of robot performance are often measured through surveys. As a more scalable and cost-effective alternative, we study the possibility of predicting people's impressions of robot behavior using non-verbal behavioral cues and machine learning techniques. To this end, we first contribute the SEAN TOGETHER Dataset consisting of observations of an interaction between a person and a mobile robot in a Virtual Reality simulation, together with impressions of robot performance provided by users on a 5-point scale. Second, we contribute analyses of how well humans and supervised learning techniques can predict perceived robot performance based on different combinations of observation types (e.g., facial, spatial, and map features). Our results show that facial expressions alone provide useful information about human impressions of robot performance; but in the navigation scenarios we tested, spatial features are the most critical piece of information for this inference task. Also, when evaluating results as binary classification (rather than multiclass classification), the F1-Score of human predictions and machine learning models more than doubles, showing that both are better at telling the directionality of robot performance than predicting exact performance ratings. Based on our findings, we provide guidelines for implementing these predictions models in real-world navigation scenarios.

show abstract