Searcharxiv⌕ Search

arXiv subjects

Maja Matarić

Publications and source records attributed to Maja Matarić.

At least 19 recordsLinked to original sources

It's My Turn! Interpretable Methods Reveal Developmental Differences from Mother-Child Conversational Dynamics

Understanding conversational interactions between mothers and children is an important challenge for computational behavioral modeling. Existing methods use traditional statistical approaches and subjective measures of conversational dynamics, without incorporating temporal or probabilistic modeling. We developed an interpretable computational framework that combines sequence analysis, information-theoretic measures, and Bayesian hierarchical modeling to characterize temporal conversational dynamics in parent-child play. We applied this framework to interactions from 94 mother-child dyads with two-year-old children. Sequence analysis revealed that dyads with lower language development (as defined by scores on the Bayley developmental test) exhibited more heterogeneous conversational organization and fewer serve-and-return exchanges than higher-development dyads, a reciprocal turn-taking pattern in which one partner initiates communication and the other responds. Mutual information analyses suggested differences in conversational predictability between lower- and higher-development dyads, although the bootstrap confidence intervals overlapped. Bayesian hierarchical models estimated temporal trajectories for maternal consecutive turns, child responses following maternal speech, and overlapping responses; HDIs showed that trajectories differed in whether they excluded 0, but there were no statistically credible differences between the Bayley groups. Overall, these results indicate that children with higher developmental scores have more organized and predictable communication styles, with more ``serve-and-return" patterns with their mothers. These results demonstrate that interpretable computational representations of conversational dynamics capture developmental differences beyond simple measures of speech quantity.

cs.HC↗

Synchrony, Affect, and Perceived Autonomy in Virtual Robot-Facilitated Anxiety Support Groups

Socially assistive robots could expand access to group-based mental health support when skilled human facilitators are unavailable. However, their effectiveness will depend on whether they foster productive interactions among participants. We examined dyadic head-movement synchrony among 96 participants across 18 virtual robot-facilitated anxiety support group sessions. A human-controlled virtual robot, via a Wizard-of-Oz (WoZ) system, facilitated all sessions. We found that stronger peak synchrony predicted lower post-session negative affect. In contrast, speakers who more often followed listeners' movements reported higher negative affect and state anxiety, demonstrating that synchrony strength and direction capture distinct interpersonal processes. Participants who believed the robot was autonomous showed more genuine synchrony and stronger peaks than those who believed it was human-operated; more negative attitudes toward robots also predicted greater synchrony. These findings support synchrony as an objective, behaviorally grounded measure for evaluating robot-facilitated groups and a potential signal for guiding future autonomous facilitators.

cs.HC↗

LazySloth: Bounded LLM-based Lazy Tree Search for Fast Long Video Comprehension

Modern vision-language models (VLMs) have shown promising results in long-video understanding due to the rich semantic information they can capture. However, most methods focus on coarse captioning of extracted image frames that are computationally inefficient and require models with large context windows. While past work has explored efficient methods through multimodal retrieval-augmented generation (RAG), they rely on lossy embeddings that lose temporal context and fine-grained detail. Few works to date have investigated how VLM-based query-relevant information retrieval can be optimized. We introduce LazySloth, an efficient tree-based search method that speeds up video comprehension and retrieval tasks 2.9-8.3x (compared to existing agentic methods) through bounded captioning of portions of the video considered irrelevant by a VLM of the video. Compared to contemporary specialized video-understanding VLMs and RAG-based methods, LazySloth achieved similar or better final task accuracy across two recent open-source base VLMs--Gemma 4 31B and Qwen3.6 27B--across four benchmarks. LazySloth reduced the gap between the base open-source model and a closed-source model, GPT-4o. Ablations showed that replacing VLM scene understanding with CLIP-based retrieval cost 8.8-19.9% in accuracy, while lazy tree construction matches eager construction at a fraction of the captioning cost. With LazySloth, we demonstrate the possibility of faster long-video comprehension without substantial loss in performance.

cs.CV↗

LLM-Powered Socially Assistive Robot-Delivered Cognitive Behavioral Therapy Exercises: an Exploratory Study with University Students

Mental health is a significant healthcare challenge, and cognitive behavioral therapy (CBT) is a widely used therapeutic method for treating anxiety and depression. However, traditional CBT often requires access to trained clinicians and can be cost-prohibitive or logistically difficult for many individuals. To address these barriers, we developed a low-cost socially assistive robot (SAR) that uses a large language model (LLM) to guide the user through interactive at-home CBT exercises. In this exploratory study, 38 university students completed CBT exercises across a 15-day period using one of three modalities: with a robot (using an LLM for dialogue), a chatbot (using the same LLM for dialogue), or traditional CBT worksheets. We measured weekly therapeutic outcomes, changes in pre-/post-session anxiety measures, and adherence to completing CBT exercises. Our findings indicate that self-reported general psychological distress significantly decreased over the study period in the robot and worksheet conditions but not in the chatbot condition. Additionally, the SAR enabled significant single-session improvements on more days than the other two conditions combined. Mixed-effects modeling further suggested that the robot and chatbot conditions better reduced post-session anxiety for those with elevated levels of anxiety. Our findings suggest that SAR-guided, LLM-powered CBT may be an effective method for supporting therapeutic progress and decreasing user anxiety immediately after completing the CBT exercise. The findings underscore the potential for combining AI-driven personalization with socially assistive robotics to create accessible, scalable, and engaging mental health interventions.

cs.RO↗

Inducing Calmness With Pocket-Sized Robotics: Reducing Movement and Heart Rate in Children through Hand-Held Tactile Interactions

Periods of heightened arousal or restlessness can interfere with children's ability to focus, self-regulation, and physically calm. Technologies that encourage embodied self-regulation through tactile interaction may provide a simple and accessible means of promoting calmness. This paper investigates how interaction with a pocket-sized tactile device influences physiological and behavioral markers of calmness in typically developing children. Building on prior work examining heart rate modulation, we present new findings on how tactile interaction affects full-body movement and postural stability. We employ a device that engages children through a hand-held rhythmic vibration-matching game, designed to focus attention and encourage stillness. Eighteen children participated in a within-subjects study that involved two conditions: with and without tactile interaction with a hand-held device, while having their heart rate and body movement recorded. Results show that the tactile game interaction reduced physiological arousal (heart rate decreased by 3.37bpm p = 0.019) and physical restlessness (overall movement decreased by 37.6% p = 0.02), with attention-related body regions showing the greatest change toward stillness (44.8% reduction in movement). These findings demonstrate that brief tactile game-like engagement with a hand-held device can down-regulate physiological activation, promoting the calm and focused states toward sustained attention and behavior regulation.

cs.RO↗

The Illusion of Intervention: Your LLM-Simulated Experiment is an Observational Study

Large language models (LLMs) show potential as simulators of human behavior, offering a scalable way to study responses to interventions. However, because LLMs are trained largely on observational data, interventions in experiments with LLM-simulated synthetic users can induce unintended shifts in latent user attributes, causing user drift where the implicit simulated population differs across treatment conditions, potentially distorting effect estimates. We formalize the confounding or selection bias that can arise due to user drift and show how intervention-dependent shifts can inflate or attenuate observed differences in user responses under intervention. To diagnose confounding, we propose using negative control outcomes--attributes that should remain invariant under intervention--to identify distribution shifts across intervention conditions, providing evidence of user drift. To mitigate drift, we study adjusting the persona specification by eliciting additional confounders, finding that targeted, setting-relevant confounders can substantially reduce bias across survey-style and multi-turn agent evaluations.

cs.CL↗

Improving through Interaction: Searching Behavioral Representation Spaces with CMA-ES-IG

Robots that interact with humans must adapt to individual users' preferences to operate effectively in human-centered environments. An intuitive and effective technique to learn non-expert users' preferences is through rankings of robot behaviors, e.g., trajectories, gestures, or voices. Existing techniques primarily focus on generating queries that optimize preference learning outcomes, such as sample efficiency or final preference estimation accuracy. However, the focus on outcome overlooks key user expectations in the process of providing these rankings, which can negatively impact users' adoption of robotic systems. This work proposes the Covariance Matrix Adaptation Evolution Strategies with Information Gain (CMA-ES-IG) algorithm. CMA-ES-IG explicitly incorporates user experience considerations into the preference learning process by suggesting perceptually distinct and informative trajectories for users to rank. We demonstrate these benefits through both simulated studies and real-robot experiments. CMA-ES-IG, compared to state-of-the-art alternatives, (1) scales more effectively to higher-dimensional preference spaces, (2) maintains computational tractability for high-dimensional problems, (3) is robust to noisy or inconsistent user feedback, and (4) is preferred by non-expert users in identifying their preferred robot behaviors. This project's code is available at github.com/interaction-lab/CMA-ES-IG

cs.RO↗

Exploring Socially Assistive Peer Mediation Robots for Teaching Conflict Resolution to Elementary School Students

In peer mediation--an approach to conflict resolution used in many K-12 schools in the United States--students help other students to resolve conflicts. For schools without peer mediation programs, socially assistive robots (SARs) may be able to provide an accessible option to practice peer mediation. We investigate how elementary school students react to a peer mediator role-play activity through an exploratory study with SARs. We conducted a small single-session between-subjects study with 12 participants. The study had two conditions, one with two robots acting as disputants, and the other without the robots and just the tablet. We found that a majority of students had positive feedback on the activity, with many students saying the peer mediation practice helped them feel better about themselves. Some said that the activity taught them how to help friends during conflict, indicating that the use of SARs for peer mediation practice is promising. We observed that participants had varying reading levels that impacted their ability to read and dictate the turns in the role-play script, an important consideration for future study design. Additionally, we found that some participants were more expressive while reading the script and throughout the activity. Although we did not find statistical differences in pre-/post-session self-perception and quiz performance between the robot and tablet conditions, we found strong correlations (p<0.05) between certain trait-related measures and learning-related measures in the robot condition, which can inform future study design for SARs for this and related contexts.

cs.HC↗

Evaluating Spoken Language as a Biomarker for Automated Screening of Cognitive Impairment

Timely and accurate assessment of cognitive impairment remains a major unmet need. Speech biomarkers offer a scalable, non-invasive, cost-effective solution for automated screening. However, the clinical utility of machine learning (ML) remains limited by interpretability and generalisability to real-world speech datasets. We evaluate explainable ML for screening of Alzheimer's disease and related dementias (ADRD) and severity prediction using benchmark DementiaBank speech (N = 291, 64% female, 69.8 (SD = 8.6) years). We validate generalisability on pilot data collected in-residence (N = 22, 59% female, 76.2 (SD = 8.0) years). To enhance clinical utility, we stratify risk for actionable triage and assess linguistic feature importance. We show that a Random Forest trained on linguistic features for ADRD detection achieves a mean sensitivity of 69.4% (95% confidence interval (CI) = 66.4-72.5) and specificity of 83.3% (78.0-88.7). On pilot data, this model yields a mean sensitivity of 70.0% (58.0-82.0) and specificity of 52.5% (39.3-65.7). For prediction of Mini-Mental State Examination (MMSE) scores, a Random Forest Regressor achieves a mean absolute MMSE error of 3.7 (3.7-3.8), with comparable performance of 3.3 (3.1-3.5) on pilot data. Risk stratification improves specificity by 13% on the test set, offering a pathway for clinical triage. Linguistic features associated with ADRD include increased use of pronouns and adverbs, greater disfluency, reduced analytical thinking, lower lexical diversity, and fewer words that reflect a psychological state of completion. Our predictive modelling shows promise for integration with conversational technology at home to monitor cognitive health and triage higher-risk individuals, enabling early screening and intervention.

cs.LG↗

Can LLMs Truly Embody Human Personality? Analyzing AI and Human Behavior Alignment in Dispute Resolution

Large language models (LLMs) are increasingly used to simulate human behavior in social settings such as legal mediation, negotiation, and dispute resolution. However, it remains unclear whether these simulations reproduce the personality-behavior patterns observed in humans. Human personality, for instance, shapes how individuals navigate social interactions, including strategic choices and behaviors in emotionally charged interactions. This raises the question: Can LLMs, when prompted with personality traits, reproduce personality-driven differences in human conflict behavior? To explore this, we introduce an evaluation framework that enables direct comparison of human-human and LLM-LLM behaviors in dispute resolution dialogues with respect to Big Five Inventory (BFI) personality traits. This framework provides a set of interpretable metrics related to strategic behavior and conflict outcomes. We additionally contribute a novel dataset creation methodology for LLM dispute resolution dialogues with matched scenarios and personality traits with respect to human conversations. Finally, we demonstrate the use of our evaluation framework with three contemporary closed-source LLMs and show significant divergences in how personality manifests in conflict across different LLMs compared to human data, challenging the assumption that personality-prompted agents can serve as reliable behavioral proxies in socially impactful applications. Our work highlights the need for psychological grounding and validation in AI simulations before real-world use.

cs.AI↗

Soft and Compliant Contact-Rich Hair Manipulation and Care

Hair care robots can help address labor shortages in elderly care while enabling those with limited mobility to maintain their hair-related identity. We present MOE-Hair, a soft robot system that performs three hair-care tasks: head patting, finger combing, and hair grasping. The system features a tendon-driven soft robot end-effector (MOE) with a wrist-mounted RGBD camera, leveraging both mechanical compliance for safety and visual force sensing through deformation. In testing with a force-sensorized mannequin head, MOE achieved comparable hair-grasping effectiveness while applying significantly less force than rigid grippers. Our novel force estimation method combines visual deformation data and tendon tensions from actuators to infer applied forces, reducing sensing errors by up to 60.1% and 20.3% compared to actuator current load-only and depth image-only baselines, respectively. A user study with 12 participants demonstrated statistically significant preferences for MOE-Hair over a baseline system in terms of comfort, effectiveness, and appropriate force application. These results demonstrate the unique advantages of soft robots in contact-rich hair-care tasks, while highlighting the importance of precise force control despite the inherent compliance of the system.

cs.RO↗

Sleepless Nights, Sugary Days: Creating Synthetic Users with Health Conditions for Realistic Coaching Agent Interactions

We present an end-to-end framework for generating synthetic users for evaluating interactive agents designed to encourage positive behavior changes, such as in health and lifestyle coaching. The synthetic users are grounded in health and lifestyle conditions, specifically sleep and diabetes management in this study, to ensure realistic interactions with the health coaching agent. Synthetic users are created in two stages: first, structured data are generated grounded in real-world health and lifestyle factors in addition to basic demographics and behavioral attributes; second, full profiles of the synthetic users are developed conditioned on the structured data. Interactions between synthetic users and the coaching agent are simulated using generative agent-based models such as Concordia, or directly by prompting a language model. Using two independently-developed agents for sleep and diabetes coaching as case studies, the validity of this framework is demonstrated by analyzing the coaching agent's understanding of the synthetic users' needs and challenges. Finally, through multiple blinded evaluations of user-coach interactions by human experts, we demonstrate that our synthetic users with health and behavioral attributes more accurately portray real human users with the same attributes, compared to generic synthetic users not grounded in such attributes. The proposed framework lays the foundation for efficient development of conversational agents through extensive, realistic, and grounded simulated interactions.

cs.LG↗

Tactile Comfort: Lowering Heart Rate Through Interactions

Children diagnosed with anxiety disorders are taught a range of strategies to navigate situations of heightened anxiety. Techniques such as deep breathing and repetition of mantras are commonly employed, as they are known to be calming and reduce elevated heart rates. Although these strategies are often effective, their successful application relies on prior training of the children for successful use when faced with challenging situations. This paper investigates a pocket-sized companion robot designed to offer a relaxation technique requiring no prior training, with a focus on immediate impact on the user's heart rate. The robot utilizes a tactile game to divert the user's attention, thereby promoting relaxation. We conducted two studies with children who were not diagnosed with anxiety: a 14-day pilot study with two children (age 8) and a main study with 18 children (ages 7-8). Both studies employed a within-subjects design and focused on measuring heart rate during tactile interaction with the robot and during non-use. Interacting with the robot was found to significantly lower the study participants' heart rate (p$<$0.01) compared to the non-use condition, indicating a consistent calming effect across all participants. These results suggest that tactile companion robots have the potential to enhance the therapeutic value of relaxation techniques.

cs.RO↗

HRIBench: Benchmarking Vision-Language Models for Real-Time Human Perception in Human-Robot Interaction

Real-time human perception is crucial for effective human-robot interaction (HRI). Large vision-language models (VLMs) offer promising generalizable perceptual capabilities but often suffer from high latency, which negatively impacts user experience and limits VLM applicability in real-world scenarios. To systematically study VLM capabilities in human perception for HRI and performance-latency trade-offs, we introduce HRIBench, a visual question-answering (VQA) benchmark designed to evaluate VLMs across a diverse set of human perceptual tasks critical for HRI. HRIBench covers five key domains: (1) non-verbal cue understanding, (2) verbal instruction understanding, (3) human-robot object relationship understanding, (4) social navigation, and (5) person identification. To construct HRIBench, we collected data from real-world HRI environments to curate questions for non-verbal cue understanding, and leveraged publicly available datasets for the remaining four domains. We curated 200 VQA questions for each domain, resulting in a total of 1000 questions for HRIBench. We then conducted a comprehensive evaluation of both state-of-the-art closed-source and open-source VLMs (N=11) on HRIBench. Our results show that, despite their generalizability, current VLMs still struggle with core perceptual capabilities essential for HRI. Moreover, none of the models within our experiments demonstrated a satisfactory performance-latency trade-off suitable for real-time deployment, underscoring the need for future research on developing smaller, low-latency VLMs with improved human perception capabilities. HRIBench and our results can be found in this Github repository: https://github.com/interaction-lab/HRIBench.

cs.RO↗

Engagement and Disclosures in LLM-Powered Cognitive Behavioral Therapy Exercises: A Factorial Design Comparing the Influence of a Robot vs. Chatbot Over Time

Many researchers are working to address the worldwide mental health crisis by developing therapeutic technologies that increase the accessibility of care, including leveraging large language model (LLM) capabilities in chatbots and socially assistive robots (SARs) used for therapeutic applications. Yet, the effects of these technologies over time remain unexplored. In this study, we use a factorial design to assess the impact of embodiment and time spent engaging in therapeutic exercises on participant disclosures. We assessed transcripts gathered from a two-week study in which 26 university student participants completed daily interactive Cognitive Behavioral Therapy (CBT) exercises in their residences using either an LLM-powered SAR or a disembodied chatbot. We evaluated the levels of active engagement and high intimacy of their disclosures (opinions, judgments, and emotions) during each session and over time. Our findings show significant interactions between time and embodiment for both outcome measures: participant engagement and intimacy increased over time in the physical robot condition, while both measures decreased in the chatbot condition.

cs.RO↗

Modeling Personalized Difficulty of Rehabilitation Exercises Using Causal Trees

Rehabilitation robots are often used in game-like interactions for rehabilitation to increase a person's motivation to complete rehabilitation exercises. By adjusting exercise difficulty for a specific user throughout the exercise interaction, robots can maximize both the user's rehabilitation outcomes and the their motivation throughout the exercise. Previous approaches have assumed exercises have generic difficulty values that apply to all users equally, however, we identified that stroke survivors have varied and unique perceptions of exercise difficulty. For example, some stroke survivors found reaching vertically more difficult than reaching farther but lower while others found reaching farther more challenging than reaching vertically. In this paper, we formulate a causal tree-based method to calculate exercise difficulty based on the user's performance. We find that this approach accurately models exercise difficulty and provides a readily interpretable model of why that exercise is difficult for both users and caretakers.

cs.RO↗

Toward Anxiety-Reducing Pocket Robots for Children

A common denominator for most therapy treatments for children who suffer from an anxiety disorder is daily practice routines to learn techniques needed to overcome anxiety. However, applying those techniques while experiencing anxiety can be highly challenging. This paper presents the design, implementation, and pilot study of a tactile hand-held pocket robot AffectaPocket, designed to work alongside therapy as a focus object to facilitate coping during an anxiety attack. The robot does not require daily practice to be used, has a small form factor, and has been designed for children 7 to 12 years old. The pocket robot works by sensing when it is being held and attempts to shift the child's focus by presenting them with a simple three-note rhythm-matching game. We conducted a pilot study of the pocket robot involving four children aged 7 to 10 years, and then a main study with 18 children aged 6 to 8 years; neither study involved children with anxiety. Both studies aimed to assess the reliability of the robot's sensor configuration, its design, and the effectiveness of the user tutorial. The results indicate that the morphology and sensor setup performed adequately and the tutorial process enabled the children to use the robot with little practice. This work demonstrates that the presented pocket robot could represent a step toward developing low-cost accessible technologies to help children suffering from anxiety disorders.

cs.RO↗

Personality Traits in Large Language Models

The advent of large language models (LLMs) has revolutionized natural language processing, enabling the generation of coherent and contextually relevant human-like text. As LLMs increasingly powerconversational agents used by the general public world-wide, the synthetic personality traits embedded in these models, by virtue of training on large amounts of human data, is becoming increasingly important. Since personality is a key factor determining the effectiveness of communication, we present a novel and comprehensive psychometrically valid and reliable methodology for administering and validating personality tests on widely-used LLMs, as well as for shaping personality in the generated text of such LLMs. Applying this method to 18 LLMs, we found: 1) personality measurements in the outputs of some LLMs under specific prompting configurations are reliable and valid; 2) evidence of reliability and validity of synthetic LLM personality is stronger for larger and instruction fine-tuned models; and 3) personality in LLM outputs can be shaped along desired dimensions to mimic specific human personality profiles. We discuss the application and ethical implications of the measurement and shaping method, in particular regarding responsible AI.

cs.CL↗