137091 research outputs found
Sort by
Multivalent immunomodulators for CAR T cell manufacturing and T cell-engager immunotherapy
No full textCancer immunotherapy, a treatment strategy in which immune cells are directed to eliminate cancer cells, represents an established therapeutic approach with a growing range of modalities and addressable disease indications. Two forms of cancer immunotherapy, chimeric antigen receptor (CAR) T cells and bi-specific T cell engagers (BiTEs), induce the clearance of malignant cells via CAR- or drug- induced T cell cytolysis, respectively. These treatments have generated promising initial responses in patients with B cell malignancies but can fail to achieve durable treatment responses following therapy. Methods to thus improve CAR T cell persistence or to enhance BiTE-mediated cytolysis of cancer cells represent an urgent and unmet clinical need. To address this need, our laboratory developed a strategy for the rapid assembly and screening of compositionally diverse libraries of IgG-conjugated, multivalent immunomodulating nanoparticles, employing it for the identification of cytokine-modified BiTEs with lytic activity comparable to FDA-approved immunotherapies. Here, we adapted this rapid assembly and screening approach to identify multivalent micro- and nano- particles that 1) selectively expanded efficacious and long-lived subsets of CAR T cells during manufacturing and 2) acted as potent BiTE immunotherapies in cell culture cytotoxicity assays incorporating B cell cancers. In this work, we studied the therapeutic impact of these new compounds 1) on the targeting, persistence, and efficacy of CAR T cell immunotherapy in mouse models of B cell malignancies and 2) as liposomal and iron oxide BiTEs for induction of malignant B cell death ex vivo. Following completion, this work presented a multivalent particle discovery platform for the identification of novel immunotherapeutics and generated new immunomodulators with the potential to improve treatment outcomes for patients with B cell malignancies.Ph.D.Biomedical Engineerin
Improving Visual Brain-Computer Interfaces Through Precise Timestamp Synchronization
No full textBrain-computer interfaces (BCIs) are communication systems where the user controls a computer with brain activities, bypassing the need for explicit motor movements. This technology is especially well-suited for those suffering from severe motor disabilities, as they cannot freely interact with computers with usual means. The usage and integration of these systems have been steadily rising over the past few decades.
BCI systems can be endogenous, where stimuli are entirely generated from the user's mental activities, or exogenous, where stimuli originates from the external environment. For BCIs designed for communication, exogenous systems have shown strong performance across various modalities such as visual, auditory, and haptics. Among these, visual evoked potential (VEP) systems, which measure brain activity in response to visual stimuli, have been shown to be the fastest and most reliable to date.
Despite significant progress, the current most performant BCI systems remain confined to controlled laboratory environments and rely on medical-grade equipment. These systems are more expensive, require bulky setups, and demand extensive user training---making them less practical for scenarios such as home care for patients or commercial use. Furthermore, they are specifically designed for environments with minimal noise, allowing for more accurate data collection. Indeed, less controlled environments, such as homes, present significant challenges for BCI systems, namely due to environmental noise, hardware constraints, and costs.
This research offers a cost-effective, open-source framework for precise timestamp synchronization, building upon OpenBCI's Cyton biosensing platform. I evaluated and characterized the performance limitations of Cyton based on its recommended canonical use. To mitigate the limitations, I proposed practical, simple-to-implement strategies and demonstrated their effectiveness. Completely solving the limitations involve custom firmware and hardware, a redesigned system architecture, and a novel, lightweight synchronization protocol uniquely leveraged for our system. I present the evaluation of an early prototype of such a system. Furthermore, a closed-loop system for verifying timestamp synchronization is introduced using external sensors. Together, these improvements enable more accurate and higher throughput data acquisition, suitable for home-based BCI applications.UndergraduateComputer Scienc
Microcat, Aquadopp, and ADCP data from the eastern mid-Atlantic ridge mooring array as part of Overturning in the Subpolar North Atlantic Program (OSNAP) from 2020-2022
No full textThe University of Miami's OSNAP (Overturning in the subpolar North Atlantic Program) is an NSF funded project that is part of the international OSNAP array put in place to measure the full depth, basin-wide overturning circulation and associated transport of heat and freshwater (www.o-snap.org). The UM Eastern MAR array consists of a series of vertical sub-surface moorings deployed along the eastern slope of the Reykjanes Ridge and across the Iceland Basin near 58°N. This dataset contains data of 2-year deployment (2020-2022) of the Deep Western Boundary Current Array in the Iceland Basin. The data sets are time series of pressure, temperature, salinity and currents and have been fully processed, calibrated and quality controlled.National Science Foundatio
Algorithms for next-generation Wi-Fi networks: multi-link operation, integrated sensing and communications, and access point mobility
No full textAs new applications with more stringent requirements and more diverse needs emerge, next-generation Wi-Fi networks are envisioned to be equipped with advanced functionalities to properly handle them. In this dissertation, we develop novel algorithms for next-generation Wi-Fi networks in terms of three advanced functionalities: multi-link operation (MLO), integrated sensing and communications (ISAC), and access point (AP) mobility. About MLO, we have proposed innovative methods of model-free dynamic traffic steering and fairness-aware dynamic link selection. About ISAC, we have designed pioneering approaches for target tracking with ISAC and with joint ISAC and MLO, respectively. About AP mobility, we have extended a prior solution with enhanced heuristic algorithms for line-of-sight discovery. Finally, we conclude this dissertation with potential future work.Ph.D.Electrical and Computer Engineerin
A Lived Experience: A Journey Through Alcohol Use Disorder
No full textThis short documentary was created as a course requirement in HTS 3086 – Sociology of Medicine and Health under the supervision of Dr. Jennifer Singh.Runtime: 05:58 minutesThis documentary explores the illness experience of recovering from Alcohol Use Disorder through Alcoholics Anonymous
Design Principles and Modelling of Microtubular Electrochemical Reactors: The Case of a Flow Battery
No full textThis thesis investigates the performance and scalability of microtubular vanadium redox flow batteries (VRFBs), addressing key factors such as ohmic resistance, electrode configuration, and material selection. A systematic approach combining experimental studies, analytical modeling, and numerical simulations provides critical insights into the challenges and opportunities for advancing microtubular reactors. Chapter 2 explores the effect of electrode porosity and configuration on ohmic losses. Experimental work demonstrates that the conductivity of the electrode and the uniformity of the current distribution are crucial for minimizing area-specific resistance (ASR). Although coaxial configurations reduce areal resistance, they may lead to higher volumetric resistance (VSR), suggesting that a quasi-coaxial configuration may be better suited for multi-tubular flow batteries. Chapter 3 presents the development of an analytical model for tubular reactors, which highlights the impact of electrode geometry and material properties on current distribution, electrode utilization, and ASR scaling. The model identifies key dimensionless parameters that govern current distribution and provides a foundation for optimizing reactor design before more complex computational methods are employed. Chapter 4 focuses on enhancing the scalability of microtubular VRFBs through material selection, specifically introducing bare copper as a promising anode material. Copper's high conductivity, stability in vanadium electrolytes, and low cost make it a strong alternative to graphite, especially for larger-scale applications. Copper demonstrates superior performance and scalability compared to graphite-based anodes. In conclusion, this thesis advances the field of electrochemical reactor design by optimizing electrode configurations, developing analytical tools, and selecting suitable materials for scalable microtubular VRFBs. The insights gained from this work contribute to the development of next-generation energy storage solutions, which are critical for the integration of renewable energy into power grids.Ph.D.Chemical and Biomolecular Engineerin
Insights into Ozone and PM2.5 Pollution: A Case Study in Spring China and Trend Analysis across the Continental United States
No full textGround-level Ozone (O3) and fine particulate matters (PM2.5) are two major pollutants, produced through complex photochemical processes involving nitrogen oxides (NOx=NO+NO2), volatile organic compounds (VOCs), and various radicals. Understanding this chemical system is crucial for effective mitigation strategies. This thesis leverages model simulations, and comprehensive ground- and satellite-based observations to gain insights into the underlying photochemistry of O3 and PM2.5 formation. This dissertation begins by exploring three observation-based pathways associated with nitrous acid (HONO) production, highlighting intrinsic relationships between NO2, particulate nitrate (pNO3) and nitric acid (HNO3). Our results reveal varying implications for O3 production. The conversion of HONO from pNO3 enhances regional O3 production, while the conversion of HONO from NO2 can reduce O3 sensitivity to NOx changes in polluted eastern China. Secondly, the comparison between satellite and ground-based multi-axis differential optical absorption spectroscopy (MAX-DOAS) measurements validates the use of satellite in assessing air pollution patterns, with better agreement observed for NO2 compared to formaldehyde (HCHO). Nonetheless, the TROPOspheric Monitoring Instrument (TROPOMI) HCHO still shows promising improvements compared to two Ozone Monitoring Instrument (OMI) products. Meanwhile, a bias ~30% between two OMI NO2 products are linked to the discrepancies between scattering weight profiles in two retrieval algorithms. In addition, regional effects, seasonal and historical trends of secondary organic carbon (SOC) across the continental United States (CONUS) through 2005-2020 are investigated using organic carbon (OC) and elemental carbon (EC) data from the Interagency Monitoring of PROtected Visual Environments (IMPROVE) network. We divided CONUS into six regions according to the correlations of OC concentrations among different sites. The regional mean secondary fractions vary from 22% to 40% and are consistent with co-located values reported by previous studies. With a consistent peak-in-summer seasonal pattern across all six regions, controlling factors for summertime SOC production for each region are investigated through a stepwise multiple linear regression. Furthermore, despite decreasing trends of anthropogenic emissions, as well as that of primary OC, significant decreasing trends in SOC are found only in eastern US in winter, and the southeast (SE) in summer. Accordingly, annual mean SOC fractions have been found to be significantly increasing except for SE. As anthropogenic emissions continue to decrease, SOC will most likely account for increasingly larger fractions of OC and PM2.5.Ph.D.Earth and Atmospheric Science
ChatHF: Collecting Rich Human Feedback from Real-time Conversations
No full textWe introduce ChatHF, an interactive annotation framework for chatbot evaluation, which integrates configurable annotation within a chat interface. ChatHF can be flexibly configured to accommodate various chatbot evaluation tasks, for example detecting offensive content, identifying incorrect or misleading information in chatbot responses, and chatbot responses that might compromise privacy. It supports post-editing of chatbot outputs and supports visual inputs, in addition to an optional voice interface. ChatHF is suitable for collection and annotation of NLP datasets, and Human-Computer Interaction studies, as demonstrated in case studies on image geolocation and assisting older adults with daily activities.M.S.Computer Scienc
Exploring Function as a Service for Emerging Latency Sensitive Applications
No full textFunction-as-a-Service (FaaS) is a promising approach to conserve resources and aid multi-tenancy in the presence of user traffic unpredictability. Emerging applications like autonomous driving and drone navigation explore the role of Edge nodes as a way to enhance perception capabilities and cope with, in many cases, limited onboard resources. Prior work on these applications have envisioned Edge node support as long running and monolithic services, and their efficacy and suitability for the FaaS paradigm is unexplored. This research effort focuses on characterizing the functions used by emergent applications and evaluating the suitability of porting such applications to be FaaS compliant.UndergraduateComputer Scienc
Resilient and sustainable transport networks: A novel decision-making framework for optimized logistics solution
No full textThis research presents a novel multi-level decision-making framework for high-speed transportation technologies for industries and small tomedium-sized enterprises. The proposed approach facilitates a stakeholder-centred selection of objectives, indicators, features, and performance criteria on the input side,and key performance indicator selection on the outputside. The framework incorporates new developments in high-speed logistics transportation, such as hyperloop technology. Advanced technologies are included, thus creating an extendedset of KPIs that current frameworks do not support, as they typically only consider existingmodes of transport. The approach utilises the Technique for Order Preference by Similarity to the Ideal Solution, which considers a variety of impacts and weight parameters. Ultimately, it aims to bridge the gap between evolving political regulations and industrial adoption within the transport sector