Search for: All records

Although generically expressing empathy is straightforward, effectively conveying empathy in specialized settings presents nuanced challenges. We present a conceptually motivated investigation into the use of figurative language and causal semantic context to facilitate targeted empathetic response generation within a specific mental health support domain, studying how these factors may be leveraged to promote improved response quality. Our approach achieves a 7.6% improvement in BLEU, a 36.7% reduction in Perplexity, and a 7.6% increase in lexical diversity (D-1 and D-2) compared to models without these signals, and human assessments show a 24.2% increase in empathy ratings. These findings provide deeper insights into grounded empathy understanding and response generation, offering a foundation for future research in this area. 

Caregiver strategy classification in pediatric rehabilitation contexts is strongly motivated by real-world clinical constraints but highly under-resourced and seldom studied in natural language processing settings. We introduce a large dataset of 4,037 caregiver strategies in this setting, a five-fold increase over the nearest contemporary dataset. These strategies are manually categorized into clinically established constructs with high agreement (𝜅=0.68-0.89). We also propose two techniques to further address identified data constraints. First, we manually supplement target task data with publicly relevant data from online child health forums. Next, we propose a novel data augmentation technique to generate synthetic caregiver strategies with high downstream task utility. Extensive experiments showcase the quality of our dataset. They also establish evidence that both the publicly available data and the synthetic strategies result in large performance gains, with relative F1 increases of 22.6% and 50.9%, respectively. 

Recent research highlights the importance of figurative language as a tool for amplifying emotional impact. In this paper, we dive deeper into this phenomenon and outline our methods for Track 1, Empathy Prediction in Conversations (CONV-dialog) and Track 2, Empathy and Emotion Prediction in Conversation Turns (CONV-turn) of the WASSA 2024 shared task. We leveraged transformer-based large language models augmented with figurative language prompts, specifically idioms, metaphors and hyperbole, that were selected and trained for each track to optimize system performance. For Track 1, we observed that a fine-tuned BERT with metaphor and hyperbole features outperformed other models on the development set. For Track 2, DeBERTa, with different combinations of figurative language prompts, performed well for different prediction tasks. Our method provides a novel framework for understanding how figurative language influences emotional perception in conversational contexts. Our system officially ranked 4th in the 1st track and 3rd in the 2nd track. 

Empathy is a social mechanism used to support and strengthen emotional connection with others, including in online communities. However, little is currently known about the nature of these online expressions, nor the particular factors that may lead to their improved detection. In this work, we study the role of a specific and complex subcategory of linguistic phenomena, figurative language, in online expressions of empathy. Our extensive experiments reveal that incorporating features regarding the use of metaphor, idiom, and hyperbole into empathy detection models improves their performance, resulting in impressive maximum F1 scores of 0.942 and 0.809 for identifying posts without and with empathy, respectively. 

Abstract Customizing participation-focused pediatric rehabilitation interventions is an important but also complex and potentially resource intensive process, which may benefit from automated and simplified steps. This research aimed at applying natural language processing to develop and identify a best performing predictive model that classifies caregiver strategies into participation-related constructs, while filtering out non-strategies. We created a dataset including 1,576 caregiver strategies obtained from 236 families of children and youth (11–17 years) with craniofacial microsomia or other childhood-onset disabilities. These strategies were annotated to four participation-related constructs and a non-strategy class. We experimented with manually created features (i.e., speech and dependency tags, predefined likely sets of words, dense lexicon features (i.e., Unified Medical Language System (UMLS) concepts)) and three classical methods (i.e., logistic regression, naïve Bayes, support vector machines (SVM)). We tested a series of binary and multinomial classification tasks applying 10-fold cross-validation on the training set (80%) to test the best performing model on the held-out test set (20%). SVM using term frequency-inverse document frequency (TF-IDF) was the best performing model for all four classification tasks, with accuracy ranging from 78.10 to 94.92% and a macro-averaged F1-score ranging from 0.58 to 0.83. Manually created features only increased model performance when filtering out non-strategies. Results suggest pipelined classification tasks (i.e., filtering out non-strategies; classification into intrinsic and extrinsic strategies; classification into participation-related constructs) for implementation into participation-focused pediatric rehabilitation interventions like Participation and Environment Measure Plus (PEM+) among caregivers who complete the Participation and Environment Measure for Children and Youth (PEM-CY). 

Health-related speech datasets are often small and varied in focus. This makes it difficult to leverage them to effectively support healthcare goals. Robust transfer of linguistic features across different datasets orbiting the same goal carries potential to address this concern. To test this hypothesis, we experiment with domain adaptation (DA) techniques on heterogeneous spoken language data to evaluate generalizability across diverse datasets for a common task: dementia detection. We find that adapted models exhibit better performance across conversational and task-oriented datasets. The feature-augmented DA method achieves a 22% increase in accuracy adapting from a conversational to task-specific dataset compared to a jointly trained baseline. This suggests promising capacity of these techniques to allow for productive use of disparate data for a complex spoken language healthcare task. 

Abstract BackgroundPractitioner and family experiences of pediatric re/habilitation can be inequitable. The Young Children’s Participation and Environment Measure (YC-PEM) is an evidence-based and promising electronic patient-reported outcome measure that was designed with and for caregivers for research and practice. This study examined historically minoritized caregivers’ responses to revised YC-PEM content modifications and their perspectives on core intelligent virtual agent functionality needed to improve its reach for equitable service design. MethodsCaregivers were recruited during a routine early intervention (EI) service visit and met five inclusion criteria: (1) were 18 + years old; (2) identified as the parent or legal guardian of a child 0–3 years old enrolled in EI services for 3 + months; (3) read, wrote, and spoke English; (4) had Internet and telephone access; and (5) identified as a parent or legal guardian of a Black, non-Hispanic child or as publicly insured. Three rounds of semi-structured cognitive interviews (55–90 min each) used videoconferencing to gather caregiver feedback on their responses to select content modifications while completing YC-PEM, and their ideas for core intelligent virtual agent functionality. Interviews were transcribed verbatim, cross-checked for accuracy, and deductively and inductively content analyzed by multiple staff in three rounds. ResultsEight Black, non-Hispanic caregivers from a single urban EI catchment and with diverse income levels (Mdn = $15,001–20,000) were enrolled, with children (M = 21.2 months,SD = 7.73) enrolled in EI. Caregivers proposed three ways to improve comprehension (clarify item wording, remove or simplify terms, add item examples). Environmental item edits prompted caregivers to share how they relate and respond to experiences with interpersonal and institutional discrimination impacting participation. Caregivers characterized three core functions of a virtual agent to strengthen YC-PEM navigation (read question aloud, visual and verbal prompts, more examples and/or definitions). ConclusionsResults indicate four ways that YC-PEM content will be modified to strengthen how providers screen for unmet participation needs and determinants to design pediatric re/habilitation services that are responsive to family priorities. Results also motivate the need for user-centered design of an intelligent virtual agent to strengthen user navigation, prior to undertaking a community-based pragmatic trial of its implementation for equitable practice. 

Background There is increased interest in using artificial intelligence (AI) to provide participation-focused pediatric re/habilitation. Existing reviews on the use of AI in participation-focused pediatric re/habilitation focus on interventions and do not screen articles based on their definition of participation. AI-based assessments may help reduce provider burden and can support operationalization of the construct under investigation. To extend knowledge of the landscape on AI use in participation-focused pediatric re/habilitation, a scoping review on AI-based participation-focused assessments is needed. Objective To understand how the construct of participation is captured and operationalized in pediatric re/habilitation using AI. Methods We conducted a scoping review of literature published in Pubmed, PsycInfo, ERIC, CINAHL, IEEE Xplore, ACM Digital Library, ProQuest Dissertation and Theses, ACL Anthology, AAAI Digital Library, and Google Scholar. Documents were screened by 2–3 independent researchers following a systematic procedure and using the following inclusion criteria: (1) focuses on capturing participation using AI; (2) includes data on children and/or youth with a congenital or acquired disability; and (3) published in English. Data from included studies were extracted [e.g., demographics, type(s) of AI used], summarized, and sorted into categories of participation-related constructs. Results Twenty one out of 3,406 documents were included. Included assessment approaches mainly captured participation through annotated observations ( n = 20; 95%), were administered in person ( n = 17; 81%), and applied machine learning ( n = 20; 95%) and computer vision ( n = 13; 62%). None integrated the child or youth perspective and only one included the caregiver perspective. All assessment approaches captured behavioral involvement, and none captured emotional or cognitive involvement or attendance. Additionally, 24% ( n = 5) of the assessment approaches captured participation-related constructs like activity competencies and 57% ( n = 12) captured aspects not included in contemporary frameworks of participation. Conclusions Main gaps for future research include lack of: (1) research reporting on common demographic factors and including samples representing the population of children and youth with a congenital or acquired disability; (2) AI-based participation assessment approaches integrating the child or youth perspective; (3) remotely administered AI-based assessment approaches capturing both child or youth attendance and involvement; and (4) AI-based assessment approaches aligning with contemporary definitions of participation. 

Background In the last decade, there has been a rapid increase in research on the use of artificial intelligence (AI) to improve child and youth participation in daily life activities, which is a key rehabilitation outcome. However, existing reviews place variable focus on participation, are narrow in scope, and are restricted to select diagnoses, hindering interpretability regarding the existing scope of AI applications that target the participation of children and youth in a pediatric rehabilitation setting. Objective The aim of this scoping review is to examine how AI is integrated into pediatric rehabilitation interventions targeting the participation of children and youth with disabilities or other diagnosed health conditions in valued activities. Methods We conducted a comprehensive literature search using established Applied Health Sciences and Computer Science databases. Two independent researchers screened and selected the studies based on a systematic procedure. Inclusion criteria were as follows: participation was an explicit study aim or outcome or the targeted focus of the AI application; AI was applied as part of the provided and tested intervention; children or youth with a disability or other diagnosed health conditions were the focus of either the study or AI application or both; and the study was published in English. Data were mapped according to the types of AI, the mode of delivery, the type of personalization, and whether the intervention addressed individual goal-setting. Results The literature search identified 3029 documents, of which 94 met the inclusion criteria. Most of the included studies used multiple applications of AI with the highest prevalence of robotics (72/94, 77%) and human-machine interaction (51/94, 54%). Regarding mode of delivery, most of the included studies described an intervention delivered in-person (84/94, 89%), and only 11% (10/94) were delivered remotely. Most interventions were tailored to groups of individuals (93/94, 99%). Only 1% (1/94) of interventions was tailored to patients’ individually reported participation needs, and only one intervention (1/94, 1%) described individual goal-setting as part of their therapy process or intervention planning. Conclusions There is an increasing amount of research on interventions using AI to target the participation of children and youth with disabilities or other diagnosed health conditions, supporting the potential of using AI in pediatric rehabilitation. On the basis of our results, 3 major gaps for further research and development were identified: a lack of remotely delivered participation-focused interventions using AI; a lack of individual goal-setting integrated in interventions; and a lack of interventions tailored to individually reported participation needs of children, youth, or families.