DSpace@RPI (Rensselaer Polytechnic Institute)
Not a member yet
6809 research outputs found
Sort by
Plasticity in molecular crystal cyclotetramethylene tetranitramine (β-hmx)
No full textMay2025School of EngineeringMolecular crystal cyclotetramethylene tetranitramine (β-HMX) is the active ingredient in widely used plastic bonded explosives. Plasticity is believed to be essential for its reaction initiation and detonation. To explore the energetic cost associated with the relative gliding of crystal planes, we calculate γ-surfaces for the most active glide planes in β-HMX, the (101) plane and the (011) plane, with pressure up to 15 GPa. Stable stacking faults are observed on both glide planes, suggesting dislocation disassociation into partials takes place. Furthermore, the γ-surface of the (101) plane indicates twinning on the (101) plane. With increasing pressure, the values of γ-surfaces increase drastically, however, the topography of γ-surfaces remains the same. Homogeneous dislocation nucleation was found to be a relevant mechanism of plastic deformation in β-HMX. In this work, we conduct atomistic simulations to investigate the conditions under which dislocations nucleate homogeneously in the (101) and (011) planes at pressures up to 20 GPa. Critical resolved shear stresses (CRSS) for dislocation nucleation are reported. The competition between the homogeneous nucleation and other mechanisms of plastic deformation shows that homogeneous nucleation is less likely to happen at pressures above 5 GPa, while at pressures below this threshold, homogeneous nucleation competes with shear localization.
Further, molecular dynamics simulations are performed to evaluate the dislocation velocity vs. resolved shear stress relation at pressures up to 20 GPa in several slip systems, which helps defining the strain rate sensitivity of the crystal. Based on this data and data from the literature, we establish a mechanism-based constitutive model for β-HMX crystals. The model captures the thermally activated and dislocation drag regimes for dislocation motion and, more importantly, the model is strongly pressure-dependent, and rate sensitive. An isotropic version of the model based of Reuss averaging is also presented. This model has the potential to be broadly applicable in the continuum modeling of HMX.
Further we study conditions under which plastic deformation in HMX becomes non-crystallographic, particularly in situations such as pore collapse under shock loading, which is considered to be a key mechanism of detonation. We observe fluidization once the applied pressure and rate are above specific thresholds, and associate this transition with the concomitant fulfillment of two conditions, one dependent on the maximum shear stress and the other dependent on the deformation rate.Ph
Exploring mast cell-mediated brain microvascular endothelial cell dysfunction in autism spectrum disorder and inflammation
No full textMay2025School of EngineeringAutism Spectrum Disorder (ASD) represents a diverse collection of neurodevelopmental disorders characterized by persistent deficits in social communication and restricted or repetitive patterns of behavior. Despite diagnoses being based on these behavioral manifestations, ASD has detrimental implications for overall health, particularly in profound ASD. The development of effective treatment remains a challenge as the etiology of this heterogenous disorder remains poorly understood. Nevertheless, pathological features such as immune dysfunction, neuroanatomical abnormalities, and neuroinflammation are consistently observed. Under neuroinflammatory conditions, alterations in blood-brain barrier (BBB) integrity are commonly noted. Compromise of the BBB – which is designed to protect the brain from the periphery and maintain a homeostatic environment – leads to increased infiltration of neurotoxins, immune cells, and inflammatory mediators. This establishes a positive feedback loop of inflammation between the central and peripheral immune systems. BBB disruption has been suggested in ASD, yet impaired BBB function and stability and its implication in the pathogenesis of ASD remains largely obscure. Here, we explore this dysfunction along with potential therapeutics in the context of immune-BBB crosstalk between brain microvascular endothelial cells (BMECs), which are the primary constituents of the BBB, and mast cells (MCs), immune cells that have implications in ASD pathogenesis. Toward exploring BMEC-MC interactions under inflammatory conditions, we first established co-cultures of primary BMECs and immortalized MCs following inflammatory stimuli, investigating their individual and combined responses and exploring potential beneficial effects of two therapeutics: rapamycin and suramin. We then developed an ASD BMEC/neural cell co-culture system using induced pluripotent stem cells (iPSCs) to investigate combined dysregulation and response to MC secretory products. Finally, we explored the therapeutic potential of quercetin and the differential response of ASD BMECs following stimulation with minor or severe MC inflammatory stimuli. Together, we observed dysregulation of neural cells and BMECs from FXS donor iPSCs. These cells exhibited greater reactivity in response to medium conditioned by activated MCs – increasing production of inflammatory cytokines and adhesion molecules. Treatment by rapamycin and suramin showed slight beneficial effects by reducing extracellular chemokines and oxidative stress in BMEC/MC co-cultures. However, quercetin treatment exhibited potent inhibition of MC activation and mitigated MC-mediated BMEC production of adhesion molecules and inflammatory cytokines/chemokines in FXS and TD. These findings provide preliminary evidence of barrier and endothelial dysfunction in FXS. Further, we report the therapeutic potential of quercetin in the context of BMEC-MC dysfunction, which may prove promising in clinical manifestations of neuroinflammation and BBB impairment.Ph
Investigating the molecular mechanisms of bacterial adaptations to pressure
No full textMay2025School of ScienceAll organisms must carefully regulate stress response pathways and the rate at which they grow and divide, and hence their size. The vast majority of microbes on Earth live in the deep biosphere, which is comprised of areas with high hydrostatic or lithostatic pressure. The molecular mechanisms underlying the adaptations of these organisms to survive in these extreme environments remain elusive. Despite this, there is great biological significance in understanding how cell growth/division and stress response pathway regulation are altered in these organisms, especially given the role of pressure in food sterilization. In this work, we investigated a pressure-adapted strain of E. coli from the perspective of both stress response and cell size, two essential properties of life. First, we determined that even pressure-adapted organisms are stressed by pressure but distinctly compared to non-adapted organisms. In particular, we demonstrated that the upregulation of the molecular chaperone, GroEL, was favored over that of the DnaK chaperone in response to pressure shock in the pressure-adapted strain, whereas the opposite was true for the non-adapted strain. We interpret this differential regulation as a consequence of the distinct functions of these two proteins. Our results also suggest that the alternative sigma factor RpoE and its anti-sigma factors may work in concert as pressure sensors. Second, we showed that the small cell size phenotype of the pressure-adapted strain is the result of the slow growth of the strain rather than an increase in the accumulation of cell division machinery. Slow growth may result from mutations in GlnA, which is implicated in the activation of the nitrogen starvation response, as well as in the RpoB subunit of RNA Polymerase. We also performed the first ever live cell imaging of FtsZ under pressure, demonstrating that the division ring formed by FtsZ is disrupted under pressure in vivo. Taken together, our results expand upon our understanding of how microbes can adapt to live in high pressure environments and survive pressure shocks.Ph
Modeling of grease lubrication behavior
No full textAugust2025School of EngineeringThe present research is aimed at building a model from basic fluid mechanics principles, deriving an appropriate modified Reynolds equation, and exploring the behavior of grease in lubrication. Greases are widely applied lubricants in industry. Despite having been used for thousands of years, they are complex substances and the lubrication mechanisms are not well understood. Due to their non-Newtonian behavior, exploration of the properties of grease is more difficult compared with oil-based Newtonian lubricants. Multiple factors can affect the performance of grease in applications. Better models are necessary to simulate the behavior of grease in working environments and to explore the properties of grease itself. The present model is built from fundamental fluid field equations including mass continuity and momentum. The geometry of the journal bearing is used, and the thin film and creeping flow conditions are applied. Multiple additional effects such as body forces, surface properties, elastohydrodynamics and energy balance (thermal effects) are not considered in starting stage of numerical studies. The Bird-Carreau model is adapted to describe the viscosity. With the numerical program, the grease flow is simulated in the geometry of the journal bearing and results predict some behaviors of grease. Pressure distribution is affected by viscosity ratio and length ratio. Eccentricity ratio, length ratio and inlet pressure have effects on load and torque. This model provides a methodology to explore additional properties of grease, and more elements could be added into the model to simulate grease in more complicated environments.Ph
An adaptive and flexible framework for convergent manufacturing with robot manipulators
No full textDecember 2024School of EngineeringRobots have become indispensable in industrial manufacturing, with easy reconfigurability,high repeatability, and the ability to operate in harsh environments. They play a crucial role
in factory production lines, executing pre-programmed motions in tasks such as packaging,
welding, and assembly. Despite their ability to perform repetitive tasks with speed and
precision, robots still face limitations in handling certain payloads (e.g., flexible or bulky
objects) or complex tasks which require significant setup, calibration, or programming efforts
(e.g., multi-robot coordination).
A significant challenge to the universal deployment of industrial robotics in advanced
manufacturing is integration and lack of robust planning. Currently, robot motion involves
manually teaching waypoints and actions through a teach pendant or pre-program motions
through commercial offline software packages. Further, the prevailing practice often involves
running robots in the open-loop mode without optimization or feedback, overlooking the
potential for improved performance and shorter cycle times with sensor-guided operation.
To address these challenges, this thesis proposes an innovative and systematic approach
to enhance industrial robotic performance and efficiency by focusing on robot interoperability,
coordination, and robustness. This is achieved by integrating various sensors across
representative manufacturing processes. This thesis explores the standardization of highlevel
robot control, robust motion planning and tracking algorithm, and a combination of
sensors with feedback to optimize the overall flexibility and performance of the robotic system.
The integration of multiple robots and sensors into a unified framework is essential for
convergent manufacturing, enhancing both performance and robustness.
Several milestones have been achieved in this thesis work, including successful robotic
manufacturing projects involving a mock assembly line, dual arm spraying, metal additive
manufacturing, and fabric handling. The ultimate goal of this work is to demonstrate a
convergent manufacturing system, comprised of multiple robots and sensors that can achieve
new or improved capabilities.Ph
Measurements of dilatational rheology of an insoluble lung surfactant using an oscillatory interfacial dialator
No full textDecember 2024School of EngineeringThis research investigates the interfacial rheology of a DPPC monolayer, the main phospholipid in lung surfactants, within a model lung flow apparatus, oscillatory interface dilator, OID. The study primarily measures interfacial velocity to explore how surfactant concentration and phase influence the behavior of a dilating monolayer, particularly regarding Marangoni flow and surface viscosity. A particle tracking velocimetry, PTV, code was utilized to capture velocity distributions across the interface. Temporal and spatial analyses revealed minimal variations in global surface measurements, highlighting the need for finely resolved data to capture subtle changes. Observations identified increased velocity variation during the compression cycle and heightened sensitivity at the cavity center with higher DPPC concentrations. Preliminary computational modeling was performed for comparative purposes, revealing limitations in simulating realistic dynamic contact angles. This study advances the understanding of surfactant dynamics at the air-liquid interface, with implications for respiratory function modeling in health and disease.M
The logic of bias: using cognitive architecture to explore interactions between cognitive abilities and decision errors
No full textDecember 2024School of Humanities, Arts, and Social SciencesViewed as cognitive imperfections, biases have been thought to be responsible for hinderinghumans from fully making use of their reasoning abilities. Several recent strains of research,
however, have begun to emphasize the positive aspect of biases. For example, PSI architec-
ture (Dörner & Güss, 2013) views these as engineered by evolution to prevent dissatisfaction,
and promote subsequent satisfaction of human needs. Crucially, PSI assumes higher skills
and reasoning capacities to enable a higher degree of introspection and cognitive flexibility,
thus alleviating the effects of biases.
Recent work by Kahan et al. (2017) called this general assumption into question: sub-
jects with higher numeracy skills were not better protected from a polarized interpretation
of statistical data, if the data contradicted their political beliefs — instead, the effect of bias
was increased. It is unclear, however, how to situate these finding within the PSI framework,
as they could be attributed to being A. a general cognitive fallacy caused to a large extent
by modulations of perceptional and attentional processes not specific to group integrity; B.
rooted in the long-term forming of stable, habituated action patterns, associated with the
subject’s beliefs; or C. an effect indeed specific to groups with strong affiliative connections.
Each of the accounts above would warrant varying revisions to the architecture.
To test the account A above, I conducted a controlled experiment examining existential
needs using thirst as a general, negative stimulus unrelated prior beliefs/experiences. The
initial results indicated that the effect reported by Kahan et al. (2017) could not be repro-
duced. This suggests that account A. — the bias being a general cognitive fallacy - does not
fit. However, the average age of the sample was quite young, as compared to Kahan’s, and
left the the possibility for account B. — the bias being rooted in long-term stable, habituated
response patterns associated with beliefs, which did not have the time to form within the
undergraduates.
viii
Testing account B required disentangling long-term beliefs from group needs. For this,
sleep/wake patterns seemed to be a good fit, as they are not strongly related to affiliation.
I conducted a study analogous to our first, where the data and background story in the
statistical task either affirmed or contradicted the subject’s sleep patterns. To control for
confounds, I asked whether these are shared by their family, peers, and/or any other larger
group they belong to. Further, the subjects sleep/wake cycles must not be determined to
a large degree by external factors. As no direct manipulation of independent variables was
required for this study, I relegated subject recruitment and data collection was relegated to
the online platform Prolific (Palan & Schitter, 2018).
As here, too, Kahan et al.’s (2017) was not observed, it
The contributions of this work are the integration of the expert bias phenomenon into
the PSI architecture by:
1. Conducting behavioral studies to explore the underpinning dynamics responsible for
the production of this effect.
2. Interpreting the findings and transcribing them onto the logic of the architecture.
3. Drafting the consequent steps necessary to initialize a revision of PSI to fit the ex-
perimentsPh
Accurate and efficient causal discovery, and causal representation learning
No full textDecember 2024School of EngineeringRecent advances in causal learning and inference have positioned causality as a key approach to addressing critical challenges in AI, including limited generalization, lack of interpretability, and fairness issues. Structural causal models (SCMs) are a foundational framework in this domain, using directed acyclic graphs (DAGs) to represent causal relationships among variables and structural equation models to quantify them. This dissertation addresses two main areas: causal discovery, which seeks to recover a unique DAG from observational data, and causal representation learning, which leverages an SCM to learn representations causally related to the target variable. We develop new theories and algorithms to overcome limitations in these areas. In causal discovery, we improve both accuracy and efficiency. To enhance accuracy, we introduce the heteroscedastic SCM (HSCM), which extends traditional SCMs by allowing heteroscedastic (variable) noise. We then propose an algorithm with identifiability guarantees to accurately learn HSCMs. Empirical results show that our method achieves high accuracy with the learned DAGs, particularly with real-world data. For efficiency, we introduce a method to learn nonlinear DAGs in a projected space, ensuring the acyclicity constraint is met without explicitly imposing it in the original space. This approach significantly reduces computation time while maintaining state-of-the-art accuracy. In causal representation learning, we address existing SCM limitations, such as limited representation scope, unaddressed latent confounders, and incomplete SCM learning. To broaden causal representations, we define an SCM that incorporates the causal Markov blanket (CMB) features of the target variable, and propose an efficient algorithm with identifiability guarantees for learning CMB representations. Results indicate significant improvement in out-of-distribution (OOD) performance. To handle latent confounders and obtain a complete SCM, we further extend the SCM by including a latent confounder, and apply the structural EM method for full SCM learning. Using the SCM, we introduce interventional inference for domain generalization and counterfactual inference for data generation, demonstrating strong OOD performance and generation of counterfactual images aligned with human interpretation.Ph
Dynamic probabilistic models for efficient and generalizable human action recognition
No full textDecember 2024School of EngineeringHuman action recognition (HAR) focuses on identifying human actions in video data. It has numerous important applications such as surveillance and smart homes. Despite its significance, HAR is challenging due to the complex dynamic patterns associated with human actions, the presence of redundant and irrelevant data in input, the high computational efficiency requirements, and the necessity to recognize both known and unknown human actions. This thesis aims to combine state-of-the-art dynamic deep models with explicit uncertainty modeling to achieve accurate, robust, and efficient HAR. First, to address the complex dynamic patterns inherent in human actions, we integrate transformer models with uncertainty modeling for complex HAR. Using the self-attention mechanism of the transformer, our model captures intricate dependencies among atomic actions. Additionally, we extend the transformer into a probabilistic transformer by treating the attention scores as random variables to capture both data and model uncertainties. Evaluations of our model on benchmarks demonstrate its superiority over existing methods. Second, to mitigate the impact of the redundant and irrelevant data on HAR, we propose an uncertainty-based spatial-temporal attention mechanism for continuous action recognition. By explicitly modeling the mutual information between the predicted labels and features using uncertainty, we generate attention masks that enable the model to prioritize high-MI features, while disregarding redundant and irrelevant ones. Evaluations on continuous action recogntion benchmarks demonstrate the effectiveness of our approach. Third, to achieve both real-time HAR along with accurate uncertainty quantification (UQ), we introduce a Bayesian Evidential Deep Learning (BEDL) framework. By combining Bayesian and evidential deep learning, BEDL employs a knowledge distillation procedure to transfer accurate UQ from a Bayesian model to a deep evidential model, which performs fast inference and precise UQ. Experiments demonstrate the accuracy and efficiency of BEDL. Finally, we tackle the challenge of recognizing both known and unknown human actions. We propose a Bayesian open-setHAR method that utilizes a deep ensemble to estimate epistemic uncertainty, enabling the distinction between known and unknown actions. Additionally, we employ optical flow to guide the model's focus on high-motion regions. Our evaluations on benchmarks reveal that our method achieves superior performance.Ph
Towards explainable and actionable bayesian deep learning
No full textDecember 2024School of EngineeringDespite significant progress in many fields, conventional deep learning models cannot effectively quantify their prediction uncertainties. They are typically overconfident in areas they do not know and they are prone to adversarial attacks and out-of-distribution inputs. To address these limitations, we propose utilizing explainable and actionable Bayesian deep learning (BDL) methods to perform accurate and efficient uncertainty quantification, identify uncertainty sources, and develop strategies to mitigate their impacts on prediction accuracy. Existing BDL methods have several shortcomings: First, they are either accurate but computationally intractable or efficient but inaccurate in uncertainty quantification. Second, they typically have limited explainability and lack an understanding of the sources of uncertainties. Finally, they often fail to mitigate uncertainties to improve model prediction performance. To address these shortcomings, this thesis focuses on three thrusts: uncertainty quantification (UQ), uncertainty attribution (UA), and uncertainty mitigation (UM). For UQ, we introduce advanced techniques to achieve a better efficiency-accuracy trade-off. The first approach enhances traditional ensemble methods by increasing component diversity, achieving state-of-the-art UQ performance. The second approach integrates an evidential neural network with Bayesian deep learning, allowing for simultaneous prediction and UQ in a single forward pass, which significantly improves computational efficiency without compromising accuracy. Additionally, we developed a gradient-based UQ method for pretrained models, enabling easy calculation of epistemic uncertainty without the need for model refinement and access to the training data. In UA, to improve explainability, we developed both gradient-based and optimization-based methods to identify problematic regions in the input that contribute to prediction uncertainty. The gradient-based method offers competitive accuracy, relaxed assumptions, and high efficiency, whereas the optimization-based method formulates UA as an optimization problem, achieving state-of-the-art performance by learning informative perturbations. Finally, for UM, we leverage insights from uncertainty attribution to develop strategies that enhance model performance. By using uncertainty attribution maps as attention mechanisms, our approach directs the model's learning toward more informative regions with low uncertainty, improving prediction accuracy and robustness.Ph