Search

IntroductionMachine learning (ML) algorithms have been heralded as promising solutions to the realization of assistive systems in digital healthcare, due to their ability to detect fine-grain patterns that are not easily perceived by humans. Yet, ML algorithms have also been critiqued for treating individuals differently based on their demography, thus propagating existing disparities. This paper explores gender and race bias in speech-based ML algorithms that detect behavioral and mental health outcomes.MethodsThis paper examines potential sources of bias in the data used to train the ML, encompassing acoustic features extracted from speech signals and associated labels, as well as in the ML decisions. The paper further examines approaches to reduce existing bias via using the features that are the least informative of one’s demographic information as the ML input, and transforming the feature space in an adversarial manner to diminish the evidence of the demographic information while retaining information about the focal behavioral and mental health state.ResultsResults are presented in two domains, the first pertaining to gender and race bias when estimating levels of anxiety, and the second pertaining to gender bias in detecting depression. Findings indicate the presence of statistically significant differences in both acoustic features and labels among demographic groups, as well as differential ML performance among groups. The statistically significant differences present in the label space are partially preserved in the ML decisions. Although variations in ML performance across demographic groups were noted, results are mixed regarding the models’ ability to accurately estimate healthcare outcomes for the sensitive groups.DiscussionThese findings underscore the necessity for careful and thoughtful design in developing ML models that are capable of maintaining crucial aspects of the data and perform effectively across all populations in digital healthcare applications.

Abstract Biosignals from wearable sensors have shown great potential for capturing environmental distress that pedestrians experience from negative stimuli (e.g., abandoned houses, poorly maintained sidewalks, graffiti, and so forth). This physiological monitoring approach in an ambulatory setting can mitigate the subjectivity and reliability concerns of traditional self-reported surveys and field audits. However, to date, most prior work has been conducted in a controlled setting and there has been little investigation into utilizing biosignals captured in real-life settings. This research examines the usability of biosignals (electrodermal activity, gait patterns, and heart rate) acquired from real-life settings to capture the environmental distress experienced by pedestrians. We collected and analyzed geocoded biosignals and self-reported stimuli information in real-life settings. Data was analyzed using spatial methods with statistical and machine learning models. Results show that the machine learning algorithm predicted location-based collective distress of pedestrians with 80% accuracy, showing statistical associations between biosignals and the self-reported stimuli. This method is expected to advance our ability to sense and react to not only built environmental issues but also urban dynamics and emergent events, which together will open valuable new opportunities to integrate human biological and physiological data streams into future built environments and/or walkability assessment applications.

IntroductionIntelligent ambulatory tracking can assist in the automatic detection of psychological and emotional states relevant to the mental health changes of professionals with high-stakes job responsibilities, such as healthcare workers. However, well-known differences in the variability of ambulatory data across individuals challenge many existing automated approaches seeking to learn a generalizable means of well-being estimation. This paper proposes a novel metric learning technique that improves the accuracy and generalizability of automated well-being estimation by reducing inter-individual variability while preserving the variability pertaining to the behavioral construct.MethodsThe metric learning technique implemented in this paper entails learning a transformed multimodal feature space from pairwise similarity information between (dis)similar samples per participant via a Siamese neural network. Improved accuracy via personalization is further achieved by considering the trait characteristics of each individual as additional input to the metric learning models, as well as individual trait base cluster criteria to group participants followed by training a metric learning model for each group.ResultsThe outcomes of the proposed models demonstrate significant improvement over the other inter-individual variability reduction and deep neural baseline methods for stress, anxiety, positive affect, and negative affect.DiscussionThis study lays the foundation for accurate estimation of psychological and emotional states in realistic and ambulatory environments leading to early diagnosis of mental health changes and enabling just-in-time adaptive interventions.

Background: The benefits of engaging in outdoor physical activity are numerous for older adults. However, previous work on outdoor monitoring of physical activities did not sufficiently identify how older adults characterize and respond to diverse elements of urban built environments, including structural characteristics, safety attributes, and aesthetics.Objective: To synthesize emerging multidisciplinary trends on the use of connected technologies to assess environmental barriers and stressors among older adults and for persons with disability.Methods: A multidisciplinary overview and literature synthesis.Results: First, we review measurement and monitoring of outdoor physical activity in community environments and during transport using wearable sensing technologies, their contextualization and using smartphone-based applications. We describe physiological responses (e.g., gait patterns, electrodermal activity, brain activity, and heart rate), stressors and physical barriers during outdoor physical activity. Second, we review the use of visual data (e.g., Google street images, Street score) and machine learning algorithms to assess physical (e.g., walkability) and emotional stressors (e.g., stress) in community environments and their impact on human perception. Third, we synthesize the challenges and limitations of using real-time smartphone-based data on driving behavior, incompatibility with software data platforms, and the potential for such data to be confounded by environmental signals in older adults. Lastly, we summarize alternative modes of transport for older adults and for persons with disability.Conclusion: Environmental design for connected technologies, interventions to promote independence and mobility, and to reduce barriers and stressors, likely requires smart connected age and disability-friendly communities and cities.

Automatic emotion recognition is the process of identifying human emotion from signals such as facial expression, speech, and text. Collecting and labeling such signals is often tedious and many times requires expert knowledge. An effective way to address challenges related to the scarcity of data and lack of human labels, is transfer learning. In this manuscript, we will describe fundamental concepts in the field of transfer learning and review work which has successfully applied transfer learning for automatic emotion recognition. We will finally discuss promising future research directions of transfer learning for improving the generalizability of automatic emotion recognition systems.

Higher education institutions are increasingly interested in innovative pedagogies that offer hands-on, project-based learning experiences in interdisciplinary settings and enable students to address real-world challenges with emerging technologies. To address this demand, a team of educators, in collaboration with community and industry partners, created an interdisciplinary and project-based service-learning program named Envisioning the Neo-traditional Development by Embracing the Autonomous Vehicles Realm (ENDEAVR). The team piloted the program during the 2019-2020 academic year. Program evaluation was conducted using both pre-post surveys and focus groups centered on the program's impact on students' critical thinking and problem-solving skills. This article summarizes key lessons, challenges and experiences from the pilot program, and shares insights on curriculum design and program implementation.