Maja J. Matarić

Allen Chang, Lauren Klein, Marcelo R. Rosales et al.

Agents must monitor their partners' affective states continuously in order to understand and engage in social interactions. However, methods for evaluating affect recognition do not account for changes in classification performance that may occur during occlusions or transitions between affective states. This paper addresses temporal patterns in affect classification performance in the context of an infant-robot interaction, where infants' affective states contribute to their ability to participate in a therapeutic leg movement activity. To support robustness to facial occlusions in video recordings, we trained infant affect recognition classifiers using both facial and body features. Next, we conducted an in-depth analysis of our best-performing models to evaluate how performance changed over time as the models encountered missing data and changing infant affect. During time windows when features were extracted with high confidence, a unimodal model trained on facial features achieved the same optimal performance as multimodal models trained on both facial and body features. However, multimodal models outperformed unimodal models when evaluated on the entire dataset. Additionally, model performance was weakest when predicting an affective state transition and improved after multiple predictions of the same affective state. These findings emphasize the benefits of incorporating body features in continuous affect recognition for infants. Our work highlights the importance of evaluating variability in model performance both over time and in the presence of missing data when applying affect recognition to social interactions.

Zhonghao Shi, Xuan Shi, Anfeng Xu et al.

Automatic speech recognition (ASR) models often experience performance degradation due to data domain shifts introduced at test time, a challenge that is further amplified for child speakers. Test-time adaptation (TTA) methods have shown great potential in bridging this domain gap. However, the use of TTA to adapt ASR models to the individual differences in each child's speech has not yet been systematically studied. In this work, we investigate the effectiveness of two widely used TTA methods-SUTA, SGEM-in adapting off-the-shelf ASR models and their fine-tuned versions for child speech recognition, with the goal of enabling continuous, unsupervised adaptation at test time. Our findings show that TTA significantly improves the performance of both off-the-shelf and fine-tuned ASR models, both on average and across individual child speakers, compared to unadapted baselines. However, while TTA helps adapt to individual variability, it may still be limited with non-linguistic child speech.

Matthew Rueben, Rhianna Lee, Thomas R. Groechel et al.

Missing significant amounts of school during K-12 education is known to put students' cognitive and social development at risk. Alternatives such as home instruction and online learning are common, but lack sufficient interaction with peers and teachers in the classroom. Mobile remote presence systems, or telepresence robots, are promising for homebound students because they provide embodiment and mobility in addition to the real-time participation offered by video conferencing technologies. Research is needed, however, for telepresence robots to meet the complex needs of homebound students participating remotely in the K-12 classroom context. We present findings from four multi-week deployments with homebound K-12 students attending classes via telepresence robots. The homebound students' experiences were documented in a total of 15 interviews and analyzed qualitatively as case studies. The homebound student participants and their deployment contexts differed from one another along multiple dimensions, and while some benefits of mobile remote attendance were enjoyed by all participants, each participant also experienced unique benefits. Some challenges with hearing, seeing, and moving the robot around the classroom warranted improvements to the design of the telepresence system. Other challenges suggested priorities for managing a classroom deployment, such as ensuring that the remote student is included in classroom activities, accountable to the teacher, and treated with respect by classmates. Based on insights from the study, we make recommendations for real-world deployment procedures in similar contexts.

Zhonghao Shi, Allison O'Connell, Zongjian Li et al.

As artificial intelligence (AI) is playing an increasingly important role in our society and global economy, AI education and literacy have become necessary components in college and K-12 education to prepare students for an AI-powered society. However, current AI curricula have not yet been made accessible and engaging enough for students and schools from all socio-economic backgrounds with different educational goals. In this work, we developed an open-source learning module for college and high school students, which allows students to build their own robot companion from the ground up. This open platform can be used to provide hands-on experience and introductory knowledge about various aspects of AI, including robotics, machine learning (ML), software engineering, and mechanical engineering. Because of the social and personal nature of a socially assistive robot companion, this module also puts a special emphasis on human-centered AI, enabling students to develop a better understanding of human-AI interaction and AI ethics through hands-on learning activities. With open-source documentation, assembling manuals and affordable materials, students from different socio-economic backgrounds can personalize their learning experience based on their individual educational goals. To evaluate the student-perceived quality of our module, we conducted a usability testing workshop with 15 college students recruited from a minority-serving institution. Our results indicate that our AI module is effective, easy-to-follow, and engaging, and it increases student interest in studying AI/ML and robotics in the future. We hope that this work will contribute toward accessible and engaging AI education in human-AI interaction for college and high school students.

Allen Chang, Matthew C. Fontaine, Serena Booth et al.

Generative models can serve as surrogates for some real data sources by creating synthetic training datasets, but in doing so they may transfer biases to downstream tasks. We focus on protecting quality and diversity when generating synthetic training datasets. We propose quality-diversity generative sampling (QDGS), a framework for sampling data uniformly across a user-defined measure space, despite the data coming from a biased generator. QDGS is a model-agnostic framework that uses prompt guidance to optimize a quality objective across measures of diversity for synthetically generated data, without fine-tuning the generative model. Using balanced synthetic datasets generated by QDGS, we first debias classifiers trained on color-biased shape datasets as a proof-of-concept. By applying QDGS to facial data synthesis, we prompt for desired semantic concepts, such as skin tone and age, to create an intersectional dataset with a combined blend of visual features. Leveraging this balanced data for training classifiers improves fairness while maintaining accuracy on facial recognition benchmarks. Code available at: https://github.com/Cylumn/qd-generative-sampling.

Emily Zhou, Mohammad Soleymani, Maja J. Matarić

Recent works have demonstrated the effectiveness of machine learning (ML) techniques in detecting anxiety and stress using physiological signals, but it is unclear whether ML models are learning physiological features specific to stress. To address this ambiguity, we evaluated the generalizability of physiological features that have been shown to be correlated with anxiety and stress to high-arousal emotions. Specifically, we examine features extracted from electrocardiogram (ECG) and electrodermal (EDA) signals from the following three datasets: Anxiety Phases Dataset (APD), Wearable Stress and Affect Detection (WESAD), and the Continuously Annotated Signals of Emotion (CASE) dataset. We aim to understand whether these features are specific to anxiety or general to other high-arousal emotions through a statistical regression analysis, in addition to a within-corpus, cross-corpus, and leave-one-corpus-out cross-validation across instances of stress and arousal. We used the following classifiers: Support Vector Machines, LightGBM, Random Forest, XGBoost, and an ensemble of the aforementioned models. We found that models trained on an arousal dataset perform relatively well on a previously unseen stress dataset, and vice versa. Our experimental results suggest that the evaluated models may be identifying emotional arousal instead of stress. This work is the first cross-corpus evaluation across stress and arousal from ECG and EDA signals, contributing new findings about the generalizability of stress detection.

Julia R. Cordero, Thomas R. Groechel, Maja J. Matarić

Study reproducibility and generalizability of results to broadly inclusive populations is crucial in any research. Previous meta-analyses in HRI have focused on the consistency of reported information from papers in various categories. However, members of the HRI community have noted that much of the information needed for reproducible and generalizable studies is not found in published papers. We address this issue by surveying the reported study metadata over the past three years (2019 through 2021) of the main proceedings of the International Conference on Human-Robot Interaction (HRI) as well as alt.HRI. Based on the analysis results, we propose a set of recommendations for the HRI community that follow the longer-standing reporting guidelines from human-computer interaction (HCI), psychology, and other fields most related to HRI. Finally, we examine three key areas for user study reproducibility: recruitment details, participant compensation, and participant gender. We find a lack of reporting within each of these study metadata categories: of the 236 studies, 139 studies failed to report recruitment method, 118 studies failed to report compensation, and 62 studies failed to report gender data. This analysis therefore provides guidance about specific types of needed reporting improvements for HRI.

Matthew Rueben, Mohammad Syed, Emily London et al.

Telepresence robots offer presence, embodiment, and mobility to remote users, making them promising options for homebound K-12 students. It is difficult, however, for robot operators to know how well they are being heard in remote and noisy classroom environments. One solution is to estimate the operator's speech intelligibility to their listeners in order to provide feedback about it to the operator. This work contributes the first evaluation of a speech intelligibility feedback system for homebound K-12 students attending class remotely. In our four long-term, in-the-wild deployments we found that students speak at different volumes instead of adjusting the robot's volume, and that detailed audio calibration and network latency feedback are needed. We also contribute the first findings about the types and frequencies of multimodal comprehension cues given to homebound students by listeners in the classroom. By annotating and categorizing over 700 cues, we found that the most common cue modalities were conversation turn timing and verbal content. Conversation turn timing cues occurred more frequently overall, whereas verbal content cues contained more information and might be the most frequent modality for negative cues. Our work provides recommendations for telepresence systems that could intervene to ensure that remote users are being heard.

Shomik Jain, Balasubramanian Thiagarajan, Zhonghao Shi et al.

Socially assistive robotics (SAR) has great potential to provide accessible, affordable, and personalized therapeutic interventions for children with autism spectrum disorders (ASD). However, human-robot interaction (HRI) methods are still limited in their ability to autonomously recognize and respond to behavioral cues, especially in atypical users and everyday settings. This work applies supervised machine learning algorithms to model user engagement in the context of long-term, in-home SAR interventions for children with ASD. Specifically, we present two types of engagement models for each user: (i) generalized models trained on data from different users; and (ii) individualized models trained on an early subset of the user's data. The models achieved approximately 90% accuracy (AUROC) for post hoc binary classification of engagement, despite the high variance in data observed across users, sessions, and engagement states. Moreover, temporal patterns in model predictions could be used to reliably initiate re-engagement actions at appropriate times. These results validate the feasibility and challenges of recognition and response to user disengagement in long-term, real-world HRI settings. The contributions of this work also inform the design of engaging and personalized HRI, especially for the ASD community.

Thomas R. Groechel, Zhonghao Shi, Roxanna Pakkar et al.

Expressivity--the use of multiple modalities to convey internal state and intent of a robot--is critical for interaction. Yet, due to cost, safety, and other constraints, many robots lack high degrees of physical expressivity. This paper explores using mixed reality to enhance a robot with limited expressivity by adding virtual arms that extend the robot's expressiveness. The arms, capable of a range of non-physically-constrained gestures, were evaluated in a between-subject study ($n=34$) where participants engaged in a mixed reality mathematics task with a socially assistive robot. The study results indicate that the virtual arms added a higher degree of perceived emotion, helpfulness, and physical presence to the robot. Users who reported a higher perceived physical presence also found the robot to have a higher degree of social presence, ease of use, usefulness, and had a positive attitude toward using the robot with mixed reality. The results also demonstrate the users' ability to distinguish the virtual gestures' valence and intent.