21793 research outputs found
Sort by
The Effect of Ligands of the β2 Adrenergic Receptor on Interleukin-2 Production in T lymphocytes
Get PDFThe immune system protects the body from infections and diseases. Helper T cells play an essential role in coordinating immune responses. In adaptive immunity, helper T cells are arguably the most important, as they are required for all adaptive immune responses. T cells activate and produce interleukin-2 (IL-2), a growth factor that causes T cells to proliferate rapidly when the body faces pathogen invasion. T cells trigger pro-inflammatory cytokine release that fights infection. However, constant T cell activation leads to an uncontrolled high amount of pro-inflammatory cytokines, causing tissue damage, resulting in chronic inflammation and, eventually, autoimmune disorders.
We previously found that nebivolol suppressed the pro-inflammatory cytokine, interleukin-17A which is implicated in the pathophysiology of inflammatory diseases and autoimmunity. Currently, there is little known about the effect of nebivolol on T cell activation regarding IL-2 production. In this thesis, I assess whether the immunomodulatory role of nebivolol extends to the IL-2 levels in T cells. My findings demonstrate that nebivolol suppressed IL-2 production in Jurkat T cells, and an inhibitory trend was observed in peripheral blood mononuclear cells (PBMCs) on activated T cells. I also found that nebivolol suppressive activity was β2-adrenergic receptor-dependent, which also implicated nuclear factor kappa B (NFκB) activity, as observed by its inhibition on phosphorylation levels.
These results are novel because there has not been a study on how nebivolol modulates IL-2 levels through the β2-adrenergic receptor. In particular, these findings add to our previous study demonstrating nebivolol immunomodulatory activity on T cells
Accelerating Graph Networks for Real-Time Physics Simulation
Get PDFPhysics-based simulations analyze interactions between physical objects, especially in areas such as terramechanics to study the interaction between soil particles. By modern standards, most computer simulations run at 60 Hz or 120 Hz refresh rates, equating to real-time step size require- ments of 16 ms or 8 ms, respectively. The discrete element method (DEM) models the interactions between the soil particles accurately, but it requires a lot of computational power. Graph Network Simulator (GNS) presents a promising alternative where it uses graph neural networks to learn and approximate the dynamics of the physical system fed into it. Although this method is faster than conventional models, the high dimensionality of the dataset means that the inference happens on a very large graph and rollouts do not occur in real-time. To accelerate the simulation, Haeri and Skonieczny (2021) developed a method that uses dimensionality reduction techniques such as principal component analysis (PCA). This method identifies the top principal components or PCA modes in the dataset that carry the highest variance. The next step is to feed these modes into the GNS along with the rigid body to train the model. Even though the reduced-order dataset results in a much smaller graph than the original dataset, the modified real-time GNS does not use algorithms such as nearest neighbours during adjacency graph construction for training and inference because the interaction between the rigid body and PCA modes may not be proximity-based. The graph fed into the network is naively fully connected. The contribution of this research is to propose a partial graph framework for the GNS to further accelerate this subspace framework without compromising the performance. To identify the redundant connections in the adjacency graph, Neural Relational Inference (NRI) is used. The NRI model, based on a variational autoencoder, uses the encoder to extract a partial graph between the PCA modes. This graph contains the most important edges that can still be used to perform inference on the GNS without any significant loss in accuracy. The effectiveness of this approach has been tested on blade cutting as well as wheel datasets to generate simulation at 60 Hz. This framework has reduced the inference time for excavation blade-driven granular flow from (0.34 - 0.48) seconds per second of simulation to (0.12 - 0.15) seconds and from 0.37 seconds to 0.18 seconds for wheel driven granular flow, achieving approximately 3x and 2x speed up respectively
Investigation on Macroscopic Mechanical Behavior of Magnetorheological Elastomers under Shear deformation, using Microscale Representative Volume Element Approach.
Get PDFMagnetorheological (MR) materials are intriguing smart materials that transform their physical characteristics in response to external magnetic fields. While MR fluids (MRFs)
are often in the spotlight due to their rapid response and field-dependent yield strength and
apparent viscosity, they grapple with issues like magnetic particle deposition and settling as well as leakage in MRF based devices. On the other hand, magnetorheological elastomers (MREs) are versatile, magnetically responsive composite materials with rubber-like qualities. By blending magnetic particles into nonmagnetic elastomeric matrices, MREs demonstrate field-dependent viscoelastic properties, such as a customizable field-dependent modulus which can not be achieved using MRFs. MREs can instantly revert to their original state when the magnetic field is withdrawn, showcasing a fascinating interplay of magnetism and material science. MREs fall into two distinct categories depending on their curing process; those formed without the influence of a magnetic field, creating isotropic MREs, and those shaped under the application of a magnetic field, yielding anisotropic MREs.
During the recent decades, the advent of finite element (FE) modeling has provided considerable benefits, by reducing the financial and time expenses forced by experimental procedures. Alongside with providing substantial accuracy, its ability to conduct parametric studies with multiple systematic parameter adjustments, ensuring greater reproducibility while simultaneously reducing the environmental impact by minimizing resource consumption and waste production are among its significant benefits. As a finite element approach, the concept of Representative Volume Element (RVE) enables predicting the material’s macroscopic behaviour from microscale modeling.
This research thesis aims at modeling the MREs’ shear behaviour under the influence of magnetic field, using the RVE concept as the modeling scheme. For this purpose, MREs with different elastomeric host matrix including silicone rubber Ecoflex 30 and Ecoflex 50 have been considered. The stress-strain characteristic of these pure silicon rubbers was first evaluated experimentally using available tensile test machine. The experimental data was then used to identify the constant parameters of the Ogden strain energy function using the least-square minimization technique. The optimized Ogden strain energy function is subsequently used in the finite element model to characterize the matrix behavior of the MREs. These, along with the mechanical and magnetic properties of Carbonyl Iron Particle (CIP), used as magnetic particles in MREs, were integrated into COMSOL Multiphysics to develop the RVE for the MREs, integrating properties of the matrix with magnetic particles. The RVE was generated in 2D and 3D configurations, for CIP volume fraction varying from 5% to 40%. Periodic Boundary Condition (PBC) was imposed on the RVE boundaries, while undergoing pure shear deformation. The results of 2D modeling suggest its ability to predict the shear deformation behaviour of isotropic MRE-RVE under varied external magnetic field. The maximum difference between theoretical and experimental shear modulus under varied magnetic field was found to be ±20%, however, the 2D model was not able to predict the MR effect at saturation with acceptable accuracy. The results from 3D modeling of isotropic MRE-RVE show a reasonably good agreement with the experiment data, with the error generally in the range of 1%- 4% for magnetic flux densities up to 0.4T. The results also suggest that the MRE-RVE with the softer silicone rubber matrix (Ecoflex 30) provides substantially higher MR effect compared with the MRE-RVE based on Ecoflex 50. Moreover, 3D isotropic MRE-RVE predicts the influence of CIP volume fraction on the shear behaviour and MR effect of MREs, comparable to experiments. All in all, the developed 3D isotropic MRE-RVE has shown the potential to accurately predict the field-dependent macroscopic mechanical properties of different MREs and thus can be used effectively to design MREs with enhanced properties without expensive experimental tests
Pull Request Abandonment in Open-Source Projects
Get PDFPull-based development is a common paradigm for contributing to and reviewing code changes in numerous open-source projects. However, a considerable amount of Pull Requests (PRs) with valid contributions are not finalized because their contributors have left the review process unfinished. Such abandoned PRs waste a considerable amount of time and effort from both their contributors and their maintainers. Furthermore, PRs that are neither progressed nor resolved, clutter the list of PRs, and eventually make it difficult for the maintainers to manage and prioritize unresolved PRs. Recognizing these challenges, this thesis aims to investigate the underlying dynamics of abandoned PRs, evaluate the helpfulness of common solutions to PR abandonment, and propose ways to mitigate PR abandonment in large open-source projects. We start by studying the characteristics of abandoned PRs, the reasons why contributors abandon their PRs, and the perspectives of project maintainers on dealing with PR abandonment. Our findings indicate that contributors and the review process play a more prominent role in PR abandonment than projects and PRs themselves. Our survey with project maintainers also indicates that Stale bot is commonly adopted by many open-source projects to deal with abandoned PRs. However, there are ongoing debates on whether using Stale bot alleviates or exacerbates PR abandonment. Therefore, in our next study, we investigate the reliance of projects on Stale bot to deal with their PR backlog, the impact of Stale bot on pull-based development, and the kind of PRs usually intervened by Stale bot. Our findings indicate that despite its benefits, Stale bot tends to further aggravate contributor abandonment. To help better mitigate PR abandonment, in our last study, we propose a machine learning approach to predict the first response latency of the maintainers and the contributor of a PR. We demonstrate the effectiveness of our approach in both project-specific and cross-project settings and also discuss the importance and impact of different features on the predicted waiting times. The awareness fostered by these predictions enables both the maintainers and the contributor to take proactive actions to mitigate potential challenges during the review process of the PR before it gets abandoned
Malware Detection and Next-Action Prediction using Learning-Based Methods
Get PDFIn this thesis, we introduce a comprehensive framework that combines natural language processing
(NLP) techniques and machine learning (ML) algorithms for the early detection and prediction
of malware activities. The core contribution of our research is the innovative application of
text classification methods, particularly Bi-LSTM neural networks and Bayesian neural networks
(BNN), to interpret application programming interface (API) call sequences as natural language
inputs. This novel approach enables us to predict upcoming malware actions, facilitating proactive
threat identification and mitigation. Our first framework employs a Bi-LSTM model to predict
the next API call, treating consecutive API calls as 2-gram and 3-gram strings. These are then
processed using a Bagging-XGBoost algorithm, enhancing the model’s ability to detect malware
presence in its early stages. The second framework advances this concept by utilizing a Bayesian
Bi-LSTM neural network. This model not only forecasts the future actions of running malware
but also quantifies the uncertainty associated with each prediction, providing a probabilistic insight
into potential malware actions. By providing the second and third most probable predictions, we
significantly improve the reliability and performance of the decision-making process. Both frameworks
are rigorously evaluated through simulations, demonstrating their effectiveness in malware
detection and action prediction. Integrating these two approaches within a single thesis represents
a significant step in applying NLP principles to cybersecurity, particularly in understanding and
countering malware threats more effectively and efficiently
Experimental and Phenomenological Examinations of Nonverbal Reassurance Seeking in Association with Obsessive-Compulsive Disorder
Get PDFIn obsessive-compulsive disorder (OCD), excessive reassurance seeking (RS) has been construed as overt direct questions and/or covert subtle statements which prompt information to allay obsessional doubts and/or fears. However, an exclusively verbal conceptualization of RS lacks theoretical justification and fails to explain the use of nonverbal actions (e.g., prolonged pauses) ostensibly to prompt reassurance following an experimental manipulation of responsibility (Leonhart & Radomsky, 2019b). Nonverbal RS (NVRS) may reinforce symptoms ‘under the radar’ of those with OCD, loved ones, and/or helping professionals. The lack of research into NVRS may limit the effectiveness of cognitive-behavioural therapy (CBT)–the treatment of choice for OCD. The present mixed-methods doctoral research was conducted to examine the manifestations and functions of NVRS in OCD. A checklist of NVRS behaviours was developed from observations and a review of the literature on RS and nonverbal communication. In Study 1, experimentally increased high ([HR] vs. decreased, low responsibility [LR]) resulted in more verbal RS (VRS) as reported by participants (N = 86) and a trained actor from whom reassurance was sought, and critically, in more NVRS according to the actor. As predicted, responsibility beliefs may have similar impacts on checking, VRS, and NVRS. Further, a significant condition × time × person interaction was observed, such that HR participants reported a smaller transfer of responsibility to the actor while LR participants reported a comparatively larger transfer to themselves following the RS opportunity. Study 2 was designed to learn if people with lived experience of seeking and providing reassurance endorsed similar and/or additional NVRS behaviour. Twelve pairs of people who met criteria for OCD and partners of their choosing (N = 24) were interviewed. Participants with OCD utilized several NVRS behaviours when given an opportunity to seek reassurance from their partners. Prominent behaviours were those which reportedly occurred during most interactions and were used at least daily to seek reassurance according to participant-partner pairs: close examination of others’ reactions, direct eye contact, pauses to allow for reassurance, and forehead wrinkling. NVRS, like VRS, may result in neutralized obsessions, but it may also have unique functions (e.g., avoided social consequences, better effectiveness). Unexpectedly, there was no clear evidence that people with OCD were unaware of their NVRS. Implications for cognitive-behavioural theories of and therapies for RS generally and NVRS specifically are discussed
Exact Approaches for a Class of Nonlinear Discrete Facility Location and Network Design Problems
No full textMany real-world problems are intrinsically mixed-integer nonlinear optimization models, which are challenging to solve due to the combinatorial complexity resulting from discrete decision variables and nonlinear functions. Mixed-integer second-order cone programming (MISOCP) is one of the promising methods to solve MINLP problems. We can describe a wide variety of real-world problems in MISOCP, such as supply chain planning, finance, network design, facility planning, and scheduling. This dissertation aims to study the MISOCP approach in facility location and network design problems where the congestion or competition settings make the problem nonlinear. We develop different types of MISOCP reformulations of nonlinear problems. We aim to find a reliable reformulation method using MISOCP to replace MINLPs which outperforms the existing solution method in the literature. We further strengthen the MISOCPs by adding valid inequality cuts.
The first paper of this thesis investigates a class of discrete stochastic facility location problems with congestion. The problem is to simultaneously determine the location and the capacity level of the facilities, as well as the allocation of customers to these facilities, aiming to minimize the combined total travel, waiting, and service times at the facilities. To model facilities, we employ spatially distributed M/G/1 queues. Incorporating waiting and service times at the facilities while determining their locations and capacities simultaneously results in a nonlinear mixed-integer programming formulation that can be computationally challenging to solve even for moderate-sized instances. We reformulate the problem as a MISOCP model. We present nine conic reformulations and polymatroid inequalities for them. Extensive computations are conducted to assess the efficacy of reformulations and determine the characteristics of a strong conic formulation.
The focus of the second paper is on the bandwidth packing problem (BPP) within telecommunication networks. The problem aims to maximize revenue by determining a set of accepted calls and routing each through the network. Excessive delays due to congestion in the network may arise if certain links are utilized close to their bandwidth capacities. The links in the network are modelled as independent M/M/1 queues. We study two variants of BPP: BPP with queueing delay cost (BPP-QDC) and BPP with queuing delay guarantees (BPP-QDG). BPP-QDC and BPP-QDG are formulated as binary integer nonlinear programs where the nonlinearity is in the objective function for BPP-QDC and the constraint for BPP-QDG. We show several ways of recasting the nonlinear BPP-QDC and BPP-QDG as MISOCPs by transforming the nonlinearity to second-order conic constraints. The reformulations are strengthened with McCormick and polymatroid inequalities. Our comprehensive computational analysis, conducted on both synthetic instances and real-world telecommunication networks, illustrates the efficacy of conic reformulations over large networks.
In the final part of this thesis, we present a branch-and-cut algorithm for a class of discrete competitive location problems (CLP), showcasing its ability to tackle extensive instances. We study our algorithms on two variants of CLP, the competitive facility location problem (CFLP) under both the MNL and gravity model and the competitive hub location problem (CHLP) under the gravity model. We present a MISOCP reformulation for this class of problems which is solved through a branch-and-cut algorithm. Our approach incorporates McCormick, polymatroid, outer approximation, and submodular inequalities. Extensive computational results demonstrate the reliability and efficacy of our algorithm in terms of CPU time, optimality gap, and the number of instances solved to optimality
Sensing Lives: Ethics, Language, and Violence on the U.S.-Mexico Border
No full textResearch on the U.S.-Mexico border commonly focuses on the institutions, social movements, or policies that shape the lives of border crossers. Here, people’s lives go unnoticed. This thesis fills that gap, looking at people’s everyday struggles to make sense of their harsh realities. The ethnographic record on the U.S.-Mexico border concentrates on the U.S. side of the border. This thesis levels the playing field by directing its attention to a catholic migrant shelter in the city of Tijuana, Mexico. At the shelter, migrants, asylum seekers, and deportees grappled with the difficulty of their realities. Storytelling was integral to making their experiences of violence thinkable. I approach people’s everyday struggles with wording their experiences through an ordinary language philosophy lens. This approach emphasizes how meaning is not secured by language but depends on how people use language to say, mean, or avoid saying, and meaning things. In this way, the use of language depends on another to pay attention to people’s words. The recognition of people’s lives through their life stories by another was important in rendering themselves accessible. This approach is as much an approach to language as to ethics and aesthetics. By wording their world, people learned something new about their lives. But this sort of knowledge, in learning to see and make differences, hinges on others’ responses. That acknowledgment was never guaranteed. At the shelter, the grueling details of people’s stories made others disavow their words. This thesis takes stock of the “problem of skepticism”, or being shut out from the world, as an ordinary dimension of people’s lives at the U.S.-Mexico border. In foregrounding people’s narratives, I show how words are vulnerable to, and depend on, our experience of the world. Words and concepts are not static screens through which we gaze upon reality, apart from it. This thesis contends that if there are no grounds for language other than its use, attention to the practices around the use of language brings us closer to the connection between us and others; to the way we engage the life of the other
Stochastic Occupant-centric Archetype Modeling for Urban Building Energy Simulation
No full textUrban Building Energy Modeling (UBEM) is vital for analyzing building energy performance, integrating detailed models of building systems, environmental factors, and occupant behavior. Despite its importance, UBEM faces challenges with uncertainty in model inputs, particularly when scaling from individual buildings to districts. This uncertainty is largely due to the stochastic nature of occupant behavior, which current models often oversimplify with fixed occupant schedules. Consequently, using the same occupant schedules for similar buildings leads to unrealistic peaks in energy demand. This thesis aims to enhance UBEM's accuracy by developing a framework for extracting and modeling realistic occupant schedules from historical data in mixed-use districts, addressing the unpredictable elements of occupant behavior and reducing the uncertainty in energy simulations. This research aims to improve UBEM through several key objectives.
Firstly, it involves creating a comprehensive database that consolidates 3D geometry models with detailed building information from various sources, effectively creating a digital twin of buildings that can be further enhanced with additional data. The second objective is to develop a standardized data model for building occupancy scheduling, tailored to different building types within UBEM. Thirdly, the research focuses on developing a data-driven method to extract representative occupant schedules, particularly related to electrical equipment usage in institutional buildings across diverse climate zones. This leads to the fourth objective: the development of a novel stochastic model using the Markov-Chain Monte Carlo (MCMC) technique, which dynamically generates occupant schedules and models energy demand at both building and urban scales. To increase efficiency and accuracy, the fifth objective is to replace the MCMC method with a Gaussian mixture model. The sixth objective integrates this stochastic model into the UBEM system to form building archetypes. The research concludes with neighborhood-level building energy simulations and model validation using real data, confirming its accuracy and real-world relevance. The study's outcome is the creation of occupant-related schedules tailored for each new building simulation, accommodating the stochastic nature of occupant behavior. The research has extensive implications, aiding in sustainable and efficient urban building design and operations, and shaping energy policy and urban planning strategies worldwide
Artist/Archaeologist: Akram Zaatari’s Memory Practice and Resisting State-Sanctioned Amnesia in the Aftermath of the Lebanese Civil Wars (1975-1990)
Get PDFCritical of the governments’ response the the Lebanese Civil Wars (1975-1990), artist, photographer, filmmaker, and archivist Akram Zaatari is among a generation of cultural producers working to oppose the politics of amnesia sponsored by the state in post-civil-war Lebanon. This thesis examines three film-based works by Zaatari and their function as a form of critical memory practice that engage with archaeological imagination and material culture to offer a space for reflection on micro-histories of the civil wars. I look closely at Red Chewing Gum (2000), In This House (2005), and Letter to Samir (2009) which focus on a particular material object and work to chip away at the state structures of forgetting. To do so, this thesis takes an oral history approach, drawing on personal interviews with the artist, in conversation with scholars of visual arts, film, archaeology, memory, and history to examine the role of material objects in engagements with the past. My analysis of Zaatari’s practice is grounded in an interdisciplinary approach, drawing largely from Laura Marks and her conception of the recollection-object and the auratic object in film, Michael Shanks and Dieter Roelstraete writing on the archaeological imagination, and the artist’s own conceptualizations of memory, excavation, and objects, often informed by Gilles Deleuze writing. I argue that Zaatari’s films display a form of memory practice which employs objects as a placeholder for past relationships, people, and time, creating an avenue to revisit an unresolved past, an archaeological practice driven by desire and care to resist forgetting