HADREB: Human Appraisals and (English) Descriptions of Robot Emotional Behaviors

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 5739–5748 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 5739 HADREB: Human Appraisals and (English) Descriptions of Robot Emotional Behaviors Josue Torres-Fonseca, Casey Kennington Boise State University 1910 W University DR, Boise, ID 83725 josuetorresfonse@u.boisestate.edu caseykennington@boisestate.edu Abstract Humans sometimes anthropomorphize everyday objects, but especially robots that have human-like qualities and that are often able to interact with and respond to humans in ways that other objects cannot. Humans especially attribute emotion to robot behaviors, partly because humans often use and interpret emotions when interacting with other humans, and they apply that capability when interacting with robots. Moreover, emotions are a fundamental part of the human language system and emotions are used as scaffolding for language learning, making them an integral part of language learning and meaning. However, there are very few datasets that explore how humans perceive the emotional states of robots and how emotional behaviors relate to human language. To address this gap we have collected HADREB, a dataset of human appraisals and English descriptions of robot emotional behaviors collected from over 30 participants. These descriptions and human emotion appraisals are collected using the Mistyrobotics Misty II and the Digital Dream Labs Cozmo (formerly Anki) robots. The dataset contains English descriptions and emotion appraisals of more than 500 descriptions and graded valence labels of 8 emotion pairs for each behavior and each robot. In this paper we describe the process of collecting and cleaning the data, give a general analysis of the data, and evaluate the usefulness of the dataset in two experiments, one using a language model to map descriptions to emotions, the other maps robot behaviors to emotions. Keywords: emotion, human-robot interaction, language learning 1. Introduction We present Human Appraisals and Descriptions of Robot Emotional Behaviors (HADREB), a dataset of human appraisals and descriptions of over 1,000 robot emotional behaviors. HADREB represents a rich re- source for investigating how humans perceive robotic behaviors relating to their emotional states.1 HADREB will be useful for incorporating emotion into lan- guage models and generating novel behaviors that al- low robots to express specific emotions. This dataset is especially useful in solving the cold-start problem of spoken dialogue systems that have no prior knowledge of language (McNeill and Kennington, 2020) because emotion exists in humans before they learn language. Emotion is also tied to the meanings of many words (Lane and Nadel, 2002); in particular, abstract words (e.g., democracy and utopia) are grounded in emotion (Vigliocco et al., 2013). Novikova et al. (2015) showed that humans anthro- pomorphize robots by attributing emotional content to robot behaviors, no matter how simple those behaviors might be, and no matter how the robot looks. Taking inspiration from Moseley et al. (2012) which posits that emotion is tied to the motor system, we design this dataset to leverage the fact that humans anthropomor- phize robot behaviors for emotional content by collect- ing human appraisals of observed robot …

Content truncated to 500 words