Virginia Tech - Wake Forest University School of Biomedical Engineering & Sciences
VTech Works (Virginia Tech)Not a member yet
122049 research outputs found
Sort by
Investigating the Use of Physiological and Behavioral Signals to Facilitate Empathic Human-AI Interaction for Daily Stress Management
No full textThis dissertation explores the design and evaluation of Empathic Large Language Models (EmLLMs) for general mental health support. EmLLMs use physiological and behavioral signals to infer users' mental states (affective and cognitive) and accordingly generate empathic messages as adaptive interventions. Three core research goals guided this work: (1) systematically reviewing state-of-the-art methods for stress and affect recognition with physiological signals and for designing physiologically adaptive systems, (2) developing and evaluating physiology-driven EmLLM prototypes that integrate stress detection with LLM-based dialogue for stress intervention, and (3) evaluating the performance and stability of multimodal LLMs using behavioral signals for emotion recognition and supportive message generation. Findings from the systematic review highlight that physiological signals provide valuable insights into stress and affect, and that systems with physiology-driven adaptation are effective at improving both user experiences and mental health interventions. Autoethnographic and pilot studies with graduate students on different prototypes of physiology-driven EmLLMs demonstrate promise for daily stress management, and expert evaluations provide further insights into refining the design of physiology-driven EmLLMs for real-world and clinical use. Performance and stability evaluations of multimodal LLMs show that multimodal behavioral inputs, including voice and facial features, enhance emotion recognition and reasoning. However, model behavior varies across modalities, underscoring the need for robust evaluation, customization strategies, and protective safeguards for mental health applications. Overall, this dissertation offers a systematic review, empirical insights, and design guidelines for developing empathic, engaging, and effective digital mental health systems.Doctor of PhilosophyManaging stress and emotional well-being is a growing challenge, especially for students and working adults. This dissertation explores how new forms of Artificial Intelligence (AI) can better understand people's emotions and support their mental health. These systems go beyond traditional digital mental health tools by using physiological signals (e.g., heart rate or skin conductance) and behavioral cues (e.g., voice or facial expressions) to estimate when someone may be stressed or overwhelmed and to respond with supportive, personalized messages. This work has three main parts. First, it reviews current scientific methods for detecting stress and emotion using physiological data and examines how technology can adapt to users' emotional states. Second, it introduces and tests several prototypes that combine physiological sensing with LLM chatbots to help graduate students reflect on and manage daily stress. Third, it evaluates how well the latest multimodal AI models can process behavioral cues to detect emotions and generate empathic responses for mental health support. Across studies, this research shows that physiological and behavioral signals can meaningfully reveal emotional patterns and that AI systems that incorporate these signals can improve user experience and emotional support. However, it also finds that AI behavior can vary across input types, underscoring the importance of careful testing, customization, and safety protections when these systems are used for mental health applications. Overall, this research provides new insights, tools, and design guidelines for creating AI systems that are not only intelligent but also sensitive, supportive, and safe to use in everyday mental health contexts
Commons: The Role of Public Space in Reinvention
No full textUrban space is, at its most basic, composed of nodes of activity and the paths that connect them. These nodes and paths form a complex urban web. The success of this fabric is determined by increasing connections and avoiding the isolation of nodes. Access to nature and pedestrian spaces form the heart of the network. These spaces must be truly free for all and allow pedestrians a place to rest and co-create. Downtown Port Chester holds a population of mainly Hispanic ethnicity, which continues to increase. The heart of Port Chester lies at the intersection of Route 1 and Westchester Ave, both forming backbones of the urban area. There have been many attempts to revitalize the downtown, but it still struggles. Hispanic restaurants in the downtown thrive, serving as the few retail spaces that continually stay occupied. This reflects a larger cultural attitude, where food is the center of community. In addition, since Port Chester is the poorest of surrounding neighborhoods, public services like libraries are chronically underfunded. This thesis proposes placing a new library at the waterfront and creating a large public park for the community. The project turns a parking lot into a community amenity, including places to rest, interact, meet, and create. The building houses both a public library and a non-profit culinary school, developing a program-hybrid designed for the future and evolution of libraries. The manipulation of light throughout the building aims to create a unique experience, especially at various times of the day, establishing a sense of place. Light is treated as a tangible thing, one that can be manipulated in order to bring out special characteristics and define space. The amount of light entering is controlled through filters of bending and passing, controlled according to the activities happening in the space and the protection from harmful UV rays. The design reinforces the idea that architecture is made up of volumes defined by light and shadow. It places light as a primary actor in shaping space. Placing this hybrid-program library on the waterfront creates a catalyst for future development of the downtown area and reimagines what the identity of Port Chester can be.Master of ArchitectureUrban spaces are shaped by concentrations of activities and the paths that connect them. Their success depends on strong connections and avoiding isolation. In Port Chester, Hispanic restaurants in the downtown thrive, due to the large Hispanic community, serving as the few retail spaces that continually stay occupied. This reflects a larger cultural attitude, where food is the center of community. In addition, since Port Chester is the poorest of the surrounding neighborhoods, public services like libraries are chronically underfunded. This thesis proposes transforming a waterfront parking lot into a public park and library, paired with a nonprofit culinary school. The manipulation of light throughout the building aims to create a unique experience, especially at various times of the day, establishing a sense of place. Light is treated as a tangible thing, one that can be manipulated in order to bring out special characteristics and define space. The design reinforces the idea that architecture is made up of volumes defined by light and shadow. Through reimagining the waterfront, the project offers a vision for revitalizing downtown and redefining Port Chester's identity
BMC Complementary Medicine and Therapies
No full textBackground: Yoga is a popular intervention demonstrating promising impacts for mental health and wellbeing. Despite growing research interest, yoga remains poorly operationalized and inconsistently described in scientific literature, hindering dissemination, rigorous evaluation, and replication. This systematic review aims to address this critical knowledge gap by examining how yoga is operationalized in recent mental health and wellbeing research.
Methods: We conducted a systematic review of literature from January 2013 to August 2024. Terms relating to yoga, mental health, wellbeing, and interventions were used to search MEDLINE, CINAHL, Embase, Emcare, PsycINFO, and Scopus. Randomized controlled trials that included yoga as the primary intervention and reported a validated measure of mental ill-health, mental wellbeing, or quality of life, were included. Inductive qualitative analyses of yoga definitions and descriptions were conducted.
Results: Of 5206 studies identified, 129 were included with exclusion primarily due to study design. Qualitative analysis resulted in a total of 1291 meaning units (MU). Yoga definitions suggest that yoga is operationalized as a practice, complementary and alternative medicine, or system (e.g., encompassing philosophy and practices) with mind-body or mind-body-spirit aspects. Components of yoga included physical such as postures, mental such as meditation, and breath.
Conclusions: This is the first systematic review to comprehensively analyze how yoga is operationalised and reported in recent experimental mental health and wellbeing research. Generally, yoga is operationalized as a mind-body or mind-body-spirit practice comprising mental, physical, and breathing components. We provide recommendations to improve the translation and implementation of yoga interventions. Trial registration This study was prospectively registered with PROSPERO (CRD42023455373). Clinical trial number: not applicable.Published versio
Beyond the Aesthetic Machine: Disney's Multiplane Camera and the Role of Special Effects Technologies in the Animation Industry
No full textThe Walt Disney Company, throughout its history, has foregrounded a commitment to technological innovation as a tool for creating stories and entertaining audiences. One of Disney's most celebrated exemplars of this commitment is the multiplane camera, a special effects device first introduced in 1937 and widely praised for its ability to instill a sense of depth in the company's earliest feature-length films, including Snow White and Pinocchio. Animation and cinema studies scholars who have studied the multiplane camera have primarily analyzed the device's aesthetic contributions.
This dissertation argues that the multiplane was far more than an aesthetic machine. Combining Cinema and Media Studies with Science and Technology Studies, it bridges the investigation of aesthetics, technology, labor, and business to explore the full lifecycle of special effects technologies in the animation industry. It approaches the multiplane camera as a networked device that both shaped and was shaped by the people, spaces, and workflows surrounding it. Using primary sources, including patents, workers' manuals, and industry publications, as well as an analysis of the aesthetic use of special effects technologies, it traces how the multiplane camera was designed, put into action, altered, and eventually constrained by shifting production practices, labor strikes, economic conditions, and consumer tastes from the 1920s to the 1950s.
By focusing on the multiplane camera, this dissertation contributes to studies of the special role of special effects and emphasizes the fragility of special effects labor as a meeting point of aesthetics, technology, and industry. It shows the negotiations required to sustain special effects technologies in practice and how once special effects become ordinary and then obsolete.Doctor of PhilosophyFor generations, The Walt Disney Company has awed audiences with films and attractions, often by using special effects technology. One system developed in the 1930s remains an icon of Disney's commitment to technologically mediated entertainment: the multiplane camera system. Disney's animators used the multiplane to separate various elements of a scene, allowing them to create a sense of three-dimensionality and depth. This special-effects technology was essential to Snow White and the Seven Dwarves and other feature films at the heart of Disney's early success. Its creators planned to make the multiplane camera a standard part of animation production and saw its future as bright.
This dissertation examines the multiplane camera in the larger network of Disney innovation from its conception, design and implementation by a new "special effects" department, use by animators in major motion pictures, centrality in marketing, role in labor relations, declining utility, and ultimate replacement by digital technologies. By exploring the multiplane camera's innovation lifecycle, I argue that the camera was more than an "aesthetic machine" that was used simply to create a specific visual effect; it became central to Disney's evolution as a creative, financial, and organizational leader. More broadly, my dissertation explores how novel innovations emerge, how they become normal technologies, and how they ultimately fade from use. It also argues for the importance of special effects production and labor as an intersection of aesthetics and technology
Development of Stochastic and Anisotropic Multiscale Ceramics Using High Aspect Ratio Sacrificial Fillers
No full textThis thesis aims to establish a straightforward and cost-effective way to form porous ceramics. There is a specific emphasis on creating anisotropic structures via a directional ice-templating technique and hard-templating using the sacrificial filler method. Cellulose nanocrystals (CNCs) aerogels have been created via a ceramic suspension of 100g/mL of CNC powder and deionized water. Samples were then frozen, in cylindrical stainless-steel molds, using a directional ice-templating method, followed by a freeze-drying process to sublimate out solvent and create a mesoporous aerogel. The porous aerogels formed were of white color and extremely lightweight, and showed various microstructural morphologies including lamellar sheeting, channels, and honeycomb structures. These samples were used as a hard-template for a chosen ceramic precursor. The precursor of choice was titanium isopropoxide (TTIP), which is a well-known titania precursor that has been reported in literature. TTIP infiltrated the aerogel structures via a one-step "wicking" procedure that induced a hydrolysis reaction between the TTIP and the hydroxyl groups within the already-formed CNC aerogel. Infiltrated aerogels were then heat-treated to 1000 °C to burnout the CNC sacrificial filler material, and densify the final porous ceramic structure. Throughout each stage of the creation of these structures, various characterization techniques were utilized to further understand morphology and chemical structure. Specific techniques include scanning electron microscopy (SEM), x-ray diffraction (XRD), thermogravimetric analysis (TGA), and Brunauer-Emmett-Teller (BET) analysis.Master of ScienceMethods for developing porous ceramic materials have gained more recent recognition for their controllability and efficiency in producing porous structures. Specifically, the method of ice-templating has been recognized as a straightforward and useful technique colloidal processing. This thesis looks at an area of materials titled cellulose nanomaterials (CNMs), and within this material class, the work detailed below utilized cellulose nanocrystals (CNCs). CNCs are desirable materials due to their renewable nature and mechanically strong properties. These materials can be utilized in various fields like biomedical devices and barrier coatings. CNCs have also been utilized in creating ceramic structures that have mesoporous and multiscale porosities. Mesoporous porosity means a material is ranging in pore size from 2-50 nm, and multiscale means that one material exhibits different sizes of porosity (macro-, meso-, and micro-) in the same structure. It is possible to use CNCs to form porous ceramics, particularly through hard-templating methods including replicas and sacrificial fillers. These methods essential create either a mirror or reverse-mirror copy of the initial structure formed. In this work, titania (TiO2) was the ceramic chosen to form. TiO2 is a white pigmented ceramic material often used in applications like coatings, paint, and thin films. This work aims to develop a method to create the mesoporous ceramics mentioned above via different colloidal processing techniques
RAPID Enabled Physics-Based Neural Networks for Predicting 3-D Fission Distributions in JSI TRIGA Reactor
No full textThe current methods for high-fidelity simulations of nuclear reactor systems are complex and computationally expensive. To reduce computation time, artificial intelligence (AI) and machine learning (ML) are being considered. Despite showing promise for solving various neutronics problems, the limited availability of high-fidelity data constrains ML applications to simpler problems or systems. This paper utilizes the RAPID code system for its effectiveness at rapidly producing large quantities of high-fidelity data. This has enabled the development of physics-based neural networks (NN) to predict 3-D fission distributions as a function of CR positions for the JSI TRIGA Mark-II research reactor. We developed a NN architecture that contains two hidden layers, 4400 neurons per hidden layer, with Leaky ReLU activation functions. This model was capable of predicting more than 99% of the fission values in the fuel elements within ±0.5% rel. diff. The model also predicted about 98% of the fission values in the fuel followers within ±10% rel. diff. It was determined that errors in the fuel follower predictions did not significantly impact calculated power peaking factors, which fell within the range of -0.39% to 0.91% rel. diff. Hyperparameter tuning and its effect on model performance is also discussed, with some comparisons to simpler ML models developed in a previous study.Accepted versionYes, full paper (Peer reviewed?
CoRR
No full textThe Adam optimizer, often used in Machine Learning for neural network training, corresponds to an underlying ordinary differential equation (ODE) in the limit of very small learning rates. This work shows that the classical Adam algorithm is a first order implicit-explicit (IMEX) Euler discretization of the underlying ODE. Employing the time discretization point of view, we propose new extensions of the Adam scheme obtained by using higher order IMEX methods to solve the ODE. Based on this approach, we derive a new optimization algorithm for neural network training that performs better than classical Adam on several regression and classification problems
npj Wireless Technology
No full textWe identify and characterize dedicated pilot symbols and other predictable elements embedded within the Starlink Ku-band downlink waveform. Exploitation of these predictable elements enables precise opportunistic positioning, navigation, and timing using compact, low-gain receivers by maximizing the signal processing gain available for signal acquisition and time-of-arrival (TOA) estimation. We develop an acquisition and demodulation framework to decode Starlink frames and disclose the explicit sequences of the edge pilots—bands of 4QAM symbols located at both edges of each Starlink channel that apparently repeat identically across all frames, beams, channels, and satellites. We further reveal that the great majority of QPSK-modulated symbols do not carry high-entropy user data but instead follow a regular tessellated structure superimposed on a constant reference template. We demonstrate that exploiting frame-level predictable elements yields a processing gain of approximately 48 dB, thereby enabling low-cost, compact receivers to extract precise TOA measurements even from low-SNR Starlink side beams.Submitted versio
