<p dir="ltr">Despite their astonishing capacity to fit data, neural networks have difficulties extrapolating beyond training data distribution. When the out-of-distribution prediction task is formalized as a counterfactual query on a causal model, the reason for their extrapolation failure is clear: neural networks learn spurious correlations in the training data rather than features that are causally related to the target label. This thesis proposes to perform a causal search over a known family of causal models to learn robust (maximally invariant) predictors for single- and multiple-environment extrapolation tasks.</p><p dir="ltr">First, I formalize the out-of-distribution task as a counterfactual query over a structural causal model. For single-environment extrapolation, I argue that symmetries of the input data are valuable for training neural networks that can extrapolate. I introduce Asymmetry learning, a new learning paradigm that is guided by the hypothesis that all (known) symmetries are mandatory even without evidence in training, unless the learner deems it inconsistent with the training data. Asymmetry learning performs a causal model search to find the simplest causal model defining a causal connection between the target labels and the symmetry transformations that affect the label. My experiments on a variety of out-of-distribution tasks on images and sequences show that proposed methods extrapolate much better than the standard neural networks.</p><p dir="ltr">Then, I consider multiple-environment out-of-distribution tasks in dynamical system forecasting that arise due to shifts in initial conditions or parameters of the dynamical system. I identify key OOD challenges in the existing deep learning and physics-informed machine learning (PIML) methods for these tasks. To mitigate these drawbacks, I combine meta-learning and causal structure discovery over a family of given structural causal models to learn the underlying dynamical system. In three simulated forecasting tasks, I show that the proposed approach is 2x to 28x more robust than the baselines.</p>

Mothers of children with autism have a higher rate of stress than mothers of neurotypical children and mothers of children with other disabilities. This impacts their family quality of life. This study aimed to show that by teaching mothers trauma-informed mindfulness-based stress reduction techniques they were able to switch their perspectives and feel that they had increased the relationships with their child with autism and increased the rating they give their family quality of life. The participants were recruited using word of mouth and through distribution of posters to autism communities following approval of the experiment through the institutional review board (IRB). The participants selected were mothers of children with autism. They interacted with the researcher through Zoom. They completed multiple measures to assess their levels of stress, well-being, mindfulness, and family quality of life. Measures included daily stress self-report, Freiburg Mindfulness Inventory-14 (FMI-14), Beach Center Family Quality of Life (FQOL), and a semi-structured interview. This was a multiple baseline study. Data analysis included visual analysis and changepoint analysis. Mothers of children with autism who utilized mindfulness, defusion, and trauma-informed problem-solving resulted in consistently decreasing levels of stress throughout the intervention. Family quality of life increased, especially in the areas of financial well-being and parenting. The mothers reported the most benefit in the practice of defusion and the least benefit from trauma-informed problem-solving. The research done in this experiment merits further study, especially in the areas of mindfulness and defusion. A larger sample size should be used to identify the benefits more closely from each phase and to identify the impact of a less homogeneous group of people. It can then be generalized to other parents of children with special needs.

Thesis advisor: Tim van Opijnen / Although vaccines and antibiotics have been historically successful in combating bacterial infections, limited vaccine coverage and the rise of antibiotic resistance emphasize the need to develop alternative, broadly effective, and/or targeted treatment strategies to reduce the health burden of bacterial infections. Rather than relying on therapeutics solely targeting the bacterial pathogen, such as standard antibiotics, therapies that simultaneously focus on host responses are emerging. In this thesis, we propose 'host-informed therapies' (HITs) in two categories: those that aid patients with fully functional immune systems and those that aid patients with perturbed immune processes, as promising alternative or adjunctive treatment strategies for bacterial infections. The host-pathogen interaction during infection is a highly dynamic process between diverse bacterial pathogens and hosts with varying degrees of susceptibility. Systems biology approaches have provided an understanding of host-pathogen parameters globally through the detection of putative biomarkers for diagnosis and identification of critical interactions to discover novel drug targets. However, there remains a gap in understanding bacterial pathogenesis in the context of designing novel host-informed therapies. Here, we use Streptococcus pneumoniae, the gram-positive pathogen responsible for the majority of bacterial respiratory tract infections worldwide, as a case study to: (1) Generate a genome-wide map of bacterial immune (complement) evasion targets to design novel host-informed therapies, (2) generate a dual host/pathogen transcriptome map to identify signatures of infection outcome, and (3) validate signatures of bacterial antibiotic tolerance in a mouse lung infection model. Overall, this work exemplifies how systems biology methods can elucidate the intricacies of bacterial pathogenesis but, more importantly, aid in the target identification, validation, and design of antibacterial host-informed therapies. / Thesis (PhD) — Boston College, 2023. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Biology.

Orbit propagation is the backbone of many problems in the space domain, such as uncertainty quantification, trajectory optimization, and guidance, navigation, and control of on orbit vehicles. Many of these techniques can rely on millions of orbit propagations, slowing computation, especially evident on low-powered satellite hardware. Past research has relied on the use of lookup tables or data streaming to enable on orbit solutions. These solutions prove inaccurate or ineffective when communication is interrupted. In this work, we introduce the use of physics-informed neural networks (PINNs) for orbit propagation to achieve fast and accurate on-board solutions, accelerated by GPU hardware solutions now available in satellite hardware. Physics-informed neural networks leverage the governing equations of motion in network training, allowing the network to optimize around the physical constraints of the system. This work leverages the use of unsupervised learning and introduces the concept of fundamental integrals of orbits to train PINNs to solve orbit problems with no knowledge of the true solution. Numerical experiments are conducted for both Earth orbits and cislunar space, being the first time a neural network integrator is implemented on flight-like hardware. The results show that the use of PINNs can decrease solution evaluation time by several order of magnitude while retaining accurate solutions to the perturbed two-body problem and the circular restricted three-body problem for deployment on spaceflight-like hardware. Implementation of these neural networks aim to reduce computational time to allow for real-time evaluation of complex algorithms on-board space vehicles.

This dissertation aims to provide critical methodological advancements for sensor fusion and physics-informed machine learning in metal additive manufacturing (MAM) to assist practitioners in detecting quality control structural anomalies. In MAM, there is an urgent need to improve knowledge of the internal layer fusion process and geometric variation occurring during the directed energy deposition processes. A core challenge lies in the cyclic heating process, which results in various structural abnormalities and deficiencies, reducing the reproducibility of manufactured components. Structural abnormalities include microstructural heterogeneities, porosity, deformation and distortion, and residual stresses. Data-driven monitoring in MAM is needed to capture process variability, but challenges arise due to the inability to capture the thermal history distribution process and structural changes below the surface due to limitations in in-situ data collection capabilities, physical domain knowledge integration, and multi-data and multi-physical data fusion. The research gaps in developing system-based generalizable artificial intelligence (AI) and machine learning (ML) to detect abnormalities are threefold. (1) Limited fusion of various types of sensor data without handcrafted selection of features. (2) There is a lack of physical domain knowledge integration for various systems, geometries, and materials. (3) It is essential to develop sensor and system integration platforms to enable a holistic view to make quality control predictions in the additive manufacturing process. In this dissertation, three studies utilize various data types and ML methodologies for predicting in-process anomalies. First, a complementary sensor fusion methodology joins thermal and ultrasonic image data capturing layer fusion and structural knowledge for layer-wise porosity segmentation. Secondly, a physics-informed data-driven methodology for joining thermal infrared image data with Goldak heat flux improves thermal history simulation and deformation detection. Lastly, a physics-informed machine learning methodology constrained by thermal physical functions utilizes in-process multi-modal monitoring data from a digital twin environment to predict distortion in the weld bead. This dissertation provides current practitioners with data-driven and physics-based interpolation methods, multi-modal sensor fusion, and anomaly detection insights trained and validated with three case studies.

Although current professional and public opinion support the right of the client to make an informed decision about entering and continuing in a psychotherapy or counseling relationship, research studying the effects of informed consent on client behaviors and attitudes in the medical, research, and mental health fields has resulted in equivocal findings. This study looked at the effects of an informed consent procedure on client behaviors and attitudes in a pro-life pregnancy counseling center where the center's primary goal is to reduce the number of abortion decisions among clients. Thirty of the center's clients (Experimental Group) were given an Informed Consent Sheet that explicitly stated the center's policies, procedures, and goals while another 30 clients {Control Group) were exposed to the center's regular procedures which did not include this Informed Consent Sheet. Results of subsequently administered questionnaires showed that there were no differences between groups regarding their stated intention to abort a potential pregnancy, nor were there any differences between groups on their attitudes toward their counselors and their counseling experience, in general.

This evidence-informed doctoral project proposes the development and implementation of a digital manual, "Safe Spaces, Stronger Futures: A Digital Guide for School-Based Occupational Therapy Practitioners Working with Children Experiencing Trauma.” In order to equip occupational therapy practitioners (OTPs) with essential trauma-informed practices, this manual offers foundational knowledge on trauma symptomatology, evidence-based strategies, interactive tools, and ongoing support through a professional online community. It aims to enhance school-based OTPs (SB-OTPs) confidence and competence in addressing trauma-related behaviors, fostering environments that support healing and growth for students with trauma histories. A comprehensive literature review was completed by the program author and addressed three key questions: (1) which interventions improve occupational performance in children who have experienced trauma, (2) what design strategies maximize adult learning, and (3) what are effective sensory supports for children with trauma histories. This project aligns with the American Occupational Therapy Association (AOTA) Occupational Therapy Practice Framework (OTPF), emphasizing mental health, sensory processing, and emotional regulation (OTPF, 2020). A pilot study has been designed to measure the program's effectiveness and to identify potential areas of improvement. This study will involve SB-OTPs from various schools in the Boston area using quantitative surveys and qualitative interviews to assess changes in practitioner knowledge, confidence, and the application of trauma-informed strategies. Expected outcomes include significant improvements in SB-OTPs' ability to implement these practices, positively impacting students affected by trauma. Dissemination plans include submitting articles to journals, presenting findings at professional conferences, and leveraging social media and online platforms to reach a broader audience. A dedicated website will ensure ongoing access to the manual and related resources, providing sustained support and professional development for SB-OTPs nationwide.

No

No / Aims and objectives; We report findings from a qualitative study to identify patient views of community hospital care. We consider how far these were in accord with the hospital staffs' views. This constituted part of a wider randomized controlled trial (RCT). The methodological challenges in seeking to identify patient satisfaction and in linking qualitative findings with trial results are explored. Design A sample of 13 patients randomized to the community hospital arm of the RCT joined the qualitative study. Official documentation from the hospital were accessed and six staff interviewed to identify assumptions underlying practice. Results Analysis of interviews identified a complex picture concerning expectations These could be classified as ideal, realistic, normative and unformed. The hospital philosophy and staff views about service delivery were closely in harmony, they delivered rehabilitation in a home-based atmosphere. The formal, or 'hard', process of rehabilitation was not well understood by patients. They were primarily concerned with 'soft' or process issues ¿ where and how care was delivered. Conclusions We identify a model of community hospital care that incorporates technical aspects of rehabilitation within a human approach that is welcomed by patients. If patients are to be able to participate in making informed decisions about care, the rationale for the activities of staff need to be more clearly explained. Recommendations are made about the appropriate scope of qualitative findings in the context of trials and about techniques to access patient views in areas where they have difficulty in expressing critical impressions.

Data from this dissertation was obtained from the Texas Behavioral Risk Factor Surveillance System (BRFSS). The Texas BRFSS questionnaire used a cross-sectional retrospective research design while asking questions about individuals' exposure to adverse childhood experiences (ACEs) and included 11 questions. The sample included 9096 individuals over the age of 18 who had exposure to at least one ACE. An ACE score was calculated for all participants and were divided into two groups (less than 4 ACEs and more than 4 ACEs) to compare whether differences in ACE score would impact participating in routine health screening or not. Additionally, whether different kinds of ACEs would impact health screening was also examined. Logistic regression was used to assess whether different kinds of ACEs impact participation in routine health screening. This study found that individuals with a history of childhood adversities including experiencing childhood emotional abuse, living in a dysfunctional household impeded them from participating in routine health screenings. It is recommended that both primary care physicians and mental health providers to use motivational interviewing while interacting with patients with ACE histories. It is also suggested that using trauma-informed care (TIC) in primary care can help patients talk about their abuse histories and utilize healthcare without any judgment.