21090 research outputs found
Sort by
Polyelectrolyte templated synthesis and formation behavior of high entropy alloys
No full textHigh entropy alloys have recently received significant attention in electrocatalysis because their unique compositional complexity can enhance both catalytic activity and long term stability. Despite this promise, there remains a lack of scalable synthesis methods that can produce nanoscale high entropy alloys with controlled and more complex morphologies. One promising strategy is to leverage the electrical double layer that forms when polyelectrolytes interact with metal salts. Polyelectrolytes can serve as effective templates by creating a locally high ion concentration along their surface, which promotes initial mixing during nanoparticle nucleation. In particular, polystyrene sulfonate can also bridge nucleating particles, allowing for the formation of more intricate, networked morphologies. The goal of this work is to investigate how polyelectrolyte concentration, polymer chain length, and different reducing agents influence the resulting catalyst composition and morphology. In addition, this study aims to provide insight into the mechanisms of nanoscale high entropy alloy formation
The Role of ICT MNCs in Climate Adaptation Through Disaster Response: Motivations, Technology, and Climate Security Implications
No full textClimate change is intensifying the frequency and severity of natural disasters with devastating effects, particularly in developing countries where vulnerabilities are amplified, while traditional disaster management and governance systems are increasingly overstretched. These climate-driven crises, which cost billions and displace millions annually, demand urgent adaptation to mitigate their catastrophic impacts on fragile societies. Multinational corporations (MNCs), especially those from the Information and Communications Technology (ICT) sector, are emerging as critical actors in disaster response, leveraging their resources and expertise to support relief and recovery efforts. This dissertation examines the role of ICT MNCs in addressing natural disasters, and explores their interventions, motivations, and potential to mitigate climate security risks by enhancing resilience in vulnerable regions. Through a multi-method approach, including bibliometric analysis, content analysis of sustainability reports, and a qualitative case study, this research reveals the growing involvement of ICT MNCs in disaster response by utilizing their technological capabilities to bridge critical gaps. However, a concerning geographical disparity is identified with declining corporate engagement in developing countries, despite their increased vulnerability. The case study of ICT MNCs' response to the 2024 Cyclone Hidaya floods in Kenya proposes the empirically grounded TEC Response framework (Triggering response–Engagement motivation–Championing technology), illustrating how corporate interventions are triggered by local presence, driven by a complex interplay of corporate social responsibility (CSR), ethical imperatives, and strategic interests, and implemented by leveraging core technological competencies. This dissertation affirms established CSR theory and contributes novel empirical insights to private governance scholarship by providing empirical evidence of the strategic and ethical dimensions of MNC involvement in disaster contexts and by highlighting their voluntary, uneven, and often unaccountable role in disaster governance, including their capacity to mitigate or inadvertently reinforce climate-induced vulnerabilities. The findings offer practical insights for policymakers and MNCs, emphasizing the importance of cross-sector collaboration, technological integration, and long-term resilience-building to enhance equitable and sustainable disaster management and climate adaptation efforts while addressing critical gaps in mitigating climate-induced vulnerabilities in fragile settings
Decoding QAnon: Building an Adaptive Alternative Reality at the Crossroads of American Conspiracism, Cultic Commodification, and Schizogenic Hyperreality
No full textQAnon has grown beyond a single conspiracy theory to become a self-perpetuating conspiracist alternative reality, one whose impact on the American political and cultural landscape will long outlive the influence of its cryptic figurehead. As bizarre as the practices of QAnon and its decoding rituals may seem, this dissertation argues that QAnon is a reflection of the techno-cultural milieu of its creation, an emergent consequence of the intersection of three key techno-cultural trends: America’s deeply entrenched cultural tradition of conspiracist narrativization, the commodification of culture under neoliberalism, and the predatory affordances of corporate media platforms optimized for the attention economy. Drawing from an array of interdisciplinary research and discursive examples drawn directly from the QAnon community, this dissertation presents a framework that can explain QAnon’s viral success within the American techno-cultural context and offer insight into the ongoing renaissance in hyper-individualistic reactionary conspiracism that QAnon has catalyzed. Only by understanding how these three trends have mutually reinforced and influenced each other can we begin to understand QAnon’s uniquely protean narrative structure and decipher the symbolic map of cultural dysfunction it represents
Integration of Muscle Pre-tension and Activation to Evaluate Neck Muscle Strain Injury Risk during Simulated Rear Impacts Using a Finite Element Neck Model
No full textPrevention of rear-impact neck injuries remains challenging for safety designers due to a lack of understanding of the tissue-level response and injury risk. Soft tissue injuries have been inferred from clinical, cadaveric, and numerical studies; however, there is a paucity of data for neck muscle injury, commonly reported as muscle pain. The goal of this study was to investigate the effect of muscle pre-tension and activation on muscle strain and injury risk resulting from low-severity rear impacts using a detailed finite element head and neck model (HNM).
The HNM was extracted from the GHBMC average stature male model and re-postured to match a volunteer study, with measured T1 kinematics applied as boundary conditions to the HNM. Three cases were simulated for three impact severities: the baseline repostured HNM, the HNM including muscle pre-tension, and the HNM with muscle pre-tension and muscle activation. The head kinematics, vertebral kinematics, muscle strains, and three neck injury criteria were calculated to assess injury risk.
The kinematic response of the neck model demonstrated an S-shaped pattern, followed by extension in the rear impact cases. The maximum kinetics, kinematics, and muscle strains occurred later in the impact during the extension phase. The distribution and magnitude of muscle strain depended on muscle pre-tension and activation, and the largest predicted strains occurred at locations associated with muscle injury reported in the literature. The HNM with muscle pre-tension and muscle activation provides a tool to assess rear impact response and could inform injury mitigation strategies in the future
Parallel Oblivious Joins using Radix Partitioning
No full textWe present parallel doubly oblivious algorithms for both non-foreign key and foreign key joins using an oblivious radix partitioning technique. Oblivious query processing enables secure execution over encrypted data when organizations outsource data to the cloud. When the cloud server processes encrypted data within hardware enclaves, the data is vulnerable to side-channel leaks caused by data-dependent memory access patterns and control flow. Our algorithms efficiently defend against these vulnerabilities by combining data partitioning with parallel execution. Specifically, we propose a doubly oblivious radix partitioning approach that divides input arrays into disjoint partitions without leaking the multiplicity of individual elements, unlike vanilla radix partitioning. This is especially important for join operations, where duplicate keys are common. To construct our join algorithm, we apply oblivious radix partitioning independently to each input table, allowing the algorithm to compare tuples only within corresponding partitions. When input tables are presorted, our oblivious join algorithm is the first to avoid combining and obliviously resorting them, yielding performance improvements over the state-of-the-art scheme, Obliviator. Beyond joins, our oblivious radix partitioning technique is a standalone primitive with applications to a broad class of problems, including oblivious aggregation and private set intersection
Democratizing and Modernizing Information Access: From Open Rerankers to Scalable RAG Evaluation
No full textModern information access increasingly relies on complex pipelines involving large language models (LLMs), fundamentally changing how users interact with information, from sophisticated multi-stage retrieval pipelines to end-to-end retrieval-augmented generation (RAG) systems. While these advancements enhance user experience, they also introduce significant challenges. The research community's growing reliance on proprietary, black-box models for key tasks like document reranking creates barriers to innovation and reproducibility (the Component Challenge). Furthermore, progress is hampered by the lack of a shared, standardized ecosystem for executing and measuring information access systems (the Benchmarking Challenge). Finally, the generative nature of RAG systems makes them fundamentally harder to evaluate than traditional systems that return document lists; new methodologies are required to assess factual accuracy and completeness in a reliable, scalable manner (the Evaluation Challenge). We argue that progress depends on the synergistic development of open, high-effectiveness system components and the reliable, scalable evaluation frameworks necessary to assess them.
This thesis addresses these challenges through a narrative arc that begins with pushing existing paradigms to their limits. But, given that frontier models dominate today's landscape, we find a pressing need for new open-source solutions. We begin by analyzing the dominant supervised ranking paradigm, developing multi-stage pipelines that demonstrated high effectiveness but also highlighted inherent complexity and cost. Subsequently, we conducted a systematic exploration of model backbones, loss functions, and negative mining strategies to squeeze effectiveness gains from supervised pointwise cross-encoders. Next, we continue with a large-scale empirical study on the newly evolving generative retrieval paradigm, which revealed its scalability limitations on large, real-world collections. This portion culminates in the final contribution to the Component Challenge: RankZephyr, an open-source 7B-parameter listwise reranker. By leveraging a carefully designed instruction distillation curriculum, RankZephyr matches and often surpasses the effectiveness of much larger proprietary models like GPT-4. It provides the community with a powerful, transparent, and accessible zero-shot reranking module, breaking the dependence on black-box systems for this critical task. All methods described have broad community adoption, and our models and evaluation frameworks continue to support ongoing research efforts across open-source IR and RAG development.
With powerful open components in hand, the focus shifts to benchmarking. To address the Benchmarking Challenge, this work introduces Ragnarök, a reusable, end-to-end RAG framework designed to standardize how retrieval-augmented generation systems are constructed and assessed. Serving as the backbone for the TREC 2024 Retrieval-Augmented Generation Track, Ragnarök provides the research community with a shared experimental platform, critical data resources, and reproducible and effective baselines. By encapsulating the full RAG pipeline — from retrieval and grounding to generation and scoring — within a single, transparent framework, TREC 2024 Retrieval-Augmented Generation Track and Ragnarök enable reproducible experimentation at scale. This not only ensures fair comparisons across diverse approaches but also establishes a foundation for cumulative progress in open-domain information access research, where previously ad hoc and non-replicable setups have often impeded reliable evaluation.
Building on this infrastructure, the thesis then directly tackles the Evaluation Challenge by introducing the AutoNuggetizer framework. This framework refactors the classic and well-studied nugget-based evaluation methodology for the modern era of LLMs. By automating the evaluation of the recall of the information nugget in RAG responses and validating the approach at scale in TREC 2024 Retrieval-Augmented Generation Track, this work provides a reliable and scalable methodology to measure the quality of generative information access systems.
In summary, this thesis contributes to the field of information access by exploring the limits of existing retrieval and ranking paradigms, developing powerful open-source components for modern information access systems, and creating the frameworks and methodologies required to benchmark and evaluate them. The contributions include a comprehensive analysis of supervised ranking and generative retrieval paradigms, an open-source state-of-the-art listwise reranker (RankZephyr), a standardized framework for RAG benchmarking (Ragnarök), and a scalable methodology for evaluating generative systems (AutoNuggetizer). Together, this thesis addresses the three core challenges identified at the outset, providing the community with both the tools to build effective systems and the methodologies to assess them rigorously. The widespread adoption of these artifacts by researchers and practitioners already underscores their tangible impact and utility in driving the field forward.
In the future, on the reranking front, we would like to build faster, more efficient rerankers that can reason over the texts and generalize to several domains. On the benchmarking front, we will expand tasks to capture "deep research" information needs that demand multi-hop reasoning and query decomposition. On the evaluation front, we hope to extend the AutoNuggetizer methodology to several tasks that go beyond web retrieval, into other domains like biomedical texts and conversational question answering
Advanced Machine Learning Techniques for Taxonomic Classification and Clustering of DNA Sequences
No full textAdvancements in genomic sequencing have exponentially increased the availability of DNA sequence data, presenting new opportunities and challenges in bioinformatics. This thesis addresses the critical need for scalable and precise computational tools to enhance taxonomic classification and clustering of DNA sequences through the application of advanced machine learning techniques. We introduce several novel algorithms designed to improve the accuracy and efficiency of these processes, focusing on both supervised and unsupervised machine learning approaches.
Firstly, we introduce "DeLUCS'', a deep learning-based method for the unsupervised clustering of DNA sequences. DeLUCS utilizes invariant information clustering to optimize the grouping of sequences without prior taxonomic information. We validate this model across multiple genomic datasets, including vertebrate mitochondrial genomes, bacterial genome segments, and viral sequences. DeLUCS significantly outperforms traditional methods such as K-Means and Gaussian Mixture Models, establishing its effectiveness for analyzing large, unlabelled DNA datasets.
Secondly, we explore the taxonomic classification of emerging astroviruses using a hybrid machine learning approach that effectively combines supervised and unsupervised techniques. Our novel methodology integrates k-mer composition analysis of whole genomes with host species data, which is crucial for accurately identifying and classifying novel and as-yet unclassified astroviruses. This approach addresses the challenges posed by genetic recombination and broad interspecies transmission that traditional host-based classifications fail to accommodate. By applying our method, we successfully proposed genus labels for 191 previously unclassified astrovirus genomes, and further identified potential cross-species infections, demonstrating the need for a revised understanding of astrovirus taxonomy.
Additionally, we present "CGRclust'', a novel method employing twin contrastive learning with convolutional neural networks (CNNs) for clustering Chaos Game Representations of DNA sequences. CGRclust leverages a unique data augmentation strategy and advanced model architecture and improves upon traditional sequence classification by avoiding the need for DNA sequence alignment or taxonomic labels. This approach demonstrates robust performance, outperforming existing methods like DeLUCS and MeShClust v3.0 with superior clustering accuracy on diverse datasets, including mitochondrial genomes and viral sequences.
Our comprehensive evaluations illustrate that these methods significantly advance the accuracy and computational efficiency of genomic data analysis. Specifically, our novel algorithms, including DeLUCS and CGRclust, leverage deep learning and contrastive learning models to refine the classification of DNA sequences, without the dependency on prior taxonomic knowledge or sequence alignment. These improvements demonstrate how machine learning can transform the field of genomic analysis, setting a new standard for taxonomic classification and providing a foundation for future explorations in evolutionary biology and biodiversity
Investigation and Enhancement of Zn-Ce Redox Flow Battery Performance Through Experimental and Modeling Studies
No full textThe transformation from energy based on fossil fuels to that based on sustainable options such as wind, solar and hydroelectric sources is crucial to reduce air/water pollution and carbon emissions. However, the production of electricity from these sustainable sources is typically intermittent in nature and can perturb the stability of the existing power grid. Redox flow batteries (RFB) have emerged as promising devices for grid-scale energy storage to stabilize power systems and improve their efficiency. Among the different types of RFBs, zinc- and cerium-based RFBs are promising for large-scale applications that require high output power density due to their low cost and high cell voltage. Motivated by its potential for future applications, this work focuses on the performance improvement of Zn-Ce RFBs through both experimental and modeling studies. Many of the findings and general ideas for RFB performance improvement are also applicable to other RFB systems and to commercial scale RFBs in real-life scenarios.
In this work, the effect of different positive supporting electrolytes on the performance of a bench-scale Zn-Ce RFB has been studied. The effectiveness of mixed methanesulfonic/sulfuric acid, mixed methanesulfonic/nitric acid and pure methanesulfonic acid has been assessed and compared. The Ce(III)/Ce(IV) reaction exhibits faster kinetics and the battery exhibits higher coulombic efficiency in the mixed 2 mol/L MSA-0.5 mol/L H2SO4 electrolyte compared to that achieved in the commonly used 4 mol/L MSA electrolyte due to lower H+ crossover and higher Ce(IV) solubility. The rate of the fade in coulombic efficiency in the mixed MSA-H2SO4 electrolyte is 0.55% per cycle over 40 charge-discharge cycles, while the fade rate is 1.26% in the case of 4 mol/L MSA. Furthermore, the positive electrode reaction is no longer the limiting half-cell reaction even at the end of long-term battery charge-discharge operation.
The effect of ion crossover on the overall Zn-Ce RFB performance has also been investigated through the measurement of the Zn(II), Ce(III), Ce(IV) and H+ concentrations on both sides of a Nafion 117 membrane during charge-discharge cycles. As much as 36% of the initial Zn(II) ions transfer from the negative to the positive electrolyte and 42.5% of the H+ in the positive electrolyte has crossed over to the negative side after 30 charge-discharge cycles. Both of these phenomena contribute to the steady fade in battery performance over the course of operation. Based on these findings, experiments aimed at reducing the concentration gradient driving crossover by intentionally adding different amounts of Zn(II) to the positive electrolyte at the outset of operation have been conducted. This approach has been shown to reduce the crossover of Zn(II) from the negative side to the positive side, improve both the battery coulombic and voltage efficiencies and reduce the decay of battery performance. Since the ion crossover phenomena is very commonly observed, this strategy to improve battery overall performance and reduce ions crossover by minimizing concentration gradient is not only applicable to similar lab-scale RFB research, but also beneficial for real-life RFB applications.
Since the positive electrode reaction becomes the limiting half-cell reaction during the course of battery operation, two strategies have been investigated to regenerate the positive electrolyte by converting the accumulated Ce(IV) ions back to Ce(III) ions. The first strategy which utilizes RuO2 as a catalyst for Ce(IV) reduction improves the voltage efficiency from 71.1% to 77.8% over 16 cycles but reduces the coulombic efficiency from 74.1% to 57.8% due to the leakage of RuO2 catalyst through the porous filter into the positive electrolyte. The method utilizing H2O2 to regenerate the positive electrolyte improves the average coulombic efficiency from 63.7% to 68.3% and the average voltage efficiency from 56.8% to 76.1% over 30 cycles. Similar battery performance and life-cycle improvement can also be expected if these electrolyte regeneration methods are applied on a commercial scale. Furthermore, the implementation of these regeneration methods should also reduce the overall operating costs since it will reduce the frequency with which electrolytes have to be replaced.
Finally, a transient 2-D model for the Zn-Ce RFB that accounts for the crossover of different electroactive species through the membrane has been developed. All three modes of transport (migration, diffusion and convection) coupled with electrode kinetics of Zn/Zn(II) and Ce(III)/Ce(IV) redox couples as well as HER and OER side reactions are included in the model. This model has been successfully validated against measurements of the evolution of the cell voltage, negative and positive electrode potentials and ion crossover during the course of 5 charge-discharge carried out in our laboratory. The validated model is then used to simulate the battery behaviour when operated under various operating conditions and using positive electrodes with different geometries. The results obtained provide useful information for the future design of Zn- or Ce-based RFBs with the aim of further improving their performance
Exploring high-level transport programming through eXpress Data Path
No full textThe transport layer is a layer from the network stack responsible for providing several essential services, such as congestion control, and lost packet detection and retransmission. Moreover, this layer is constantly changing to satisfy the new demands present in the network. This includes the increase in traffic, the leveraging of new hardware, and the release of new workloads.
However, implementing these protocols is a considerably challenging task due to several reasons. First, the available documentation about these algorithms is written in natural language and can be long, which may result in misinterpretations, and consequently, incorrect implementations. Second, while developing transport algorithms, programmers must deeply study the environment where the protocol will run to find the best-suiting built-in helper functions and data structures available for their implementations. For this reason, the protocols are usually implemented as large sections of optimized code that are challenging to navigate, read, and modify. Third, since the transport layer is located between two other layers, namely application and network, transport algorithms must handle the details of these interactions, which can depend on the execution environment.
In this thesis, we introduce Modular Transport Programming (MTP). This event-driven framework allows the declaration of transport algorithms with a high-level, yet, precise language. MTP abstracts low-level details by providing a set of high-level built-in functions and constructs focused solely on transport programming. It also abstracts the interactions between the transport layer and its neighbours by having interfaces responsible for such communication. Moreover, it improves the readability of MTP code by having a modularized design that clearly shows how each module interacts with the other and how the events are processed.
To explore the feasibility and effectiveness of MTP, we chose to implement transport protocols in the Linux Kernel. The Linux Kernel and its transport layer implementations are some of the most notable examples that highlight the challenges stated above. These implementations can require thousands of lines of code and hundreds of functions, which are difficult to navigate due to the low-level C used to write their codes. Additionally, transport protocol implementations in the Kernel are tightly linked with the neighbouring layers of the network stack, utilizing complex data structures for packet processing and socket interface. Thus, implementing protocols from the start or modifying already existing ones can be a considerably difficult task.
Nevertheless, instead of directly implementing protocols to the Linux Kernel, we opted to use the eXpress Data Path (XDP) framework as our first step. This framework allows a simple and safe loading of new user-defined code to the Linux Kernel. Moreover, several characteristics of the framework are ideal for the implementation of transport layer protocols, such as its loading point being in the Network Interface Card (NIC) driver and allowing Kernel bypass. Thus, this thesis also introduces a back-end developed in the XDP framework, to which the MTP code can be compiled using our code generator. During the development of this back-end, we explored XDP in depth to make wise design decisions. This includes utilizing a set of data structures best suited for our model, making the most of the helper functions available in its libraries, and bypassing the limitations of this framework.
Finally, we show that it is possible to specify the logic of transport protocols in a high-level language and translate it to XDP. We implemented two protocols, TCP and RoCEv2, in the MTP language and successfully translated them with our compiler to the XDP back-end. Moreover, between two servers connected by 100Gbps links, the TCP implementation presented 80Gbps of throughput with 16 cores processing packets. Meanwhile, RoCEv2 translation is functional but still needs further optimizations to reach its expected performance. Lastly, we evaluate the strengths and weaknesses of XDP for transport programming
Process Parameter Optimization for Crack Mitigation in CM247LC processed by Electron Beam Powder Bed Fusion
No full textPowder bed fusion is a class of additive manufacturing (AM) technology capable of fabricating complex geometrical designs and empowering users with a variety of benefits. Two main examples are laser powder bed fusion (LPBF) and electron beam powder bed fusion (EB-PBF) known as electron beam melting (EBM). Both techniques follow a layer-by-layer fabrication manner whereby each layer is melted according to individual cross-sectional slices of a computer aided design model. Despite its escalating enticement and prominence in both research and industry, the application requires significant research to understand the interplay between process parameters and materials. An example of a group of alloys are nickel superalloys, which can be further classified as weldable and non-weldable nickel superalloys. CM247LC falls in the latter and is associated with a higher volume content of γ' precipitation strengthening phase. The greater γ' (gamma prime) content enhances the mechanical properties for high temperature applications, but also paradoxically contributes to its proclivity for cracking.
The processability of CM247LC via LPBF has been investigated by different researchers from academia and industry. A number of traditional and creative strategies were explored to minimize the crack susceptibility of CM247LC but this was at the cost of undesired compromises. Conversely, the differences in the deposition process and conditions primes the less explored EBM to be an appealing processing alternative. The current work explores the processability of CM247LC by EBM with the desired target of mitigation the material susceptibility towards cracking. The process optimization requires a fundamental understanding of process parameters and the solidification process of the deposited material. This study follows a systematic ground up approach to address the knowledge gap. This started with the basic deposition unit of a single-track and later progressed with multi-tracks.
An experimental single-track study, structured within a design of experiment (DoE) framework, was conducted to isolate the primary EBM process parameters and evaluate their effects on the track stability, microstructure, and cracking behaviour. A process map was developed to identify parameter combinations that yielded coherent and uniform tracks. Relationships among EBM parameters, melt pool morphology, grain microstructure, and cracking tendencies were established, revealing specific track conditions and inherent characteristics that mitigate cracking.
The lack of an integrated characterization system within the build chamber complicates direct analysis of solidification conditions for these single-tracks. To address this, a simplified thermal model by finite element (FEM) method was developed to provide an alternate approach to study the solidification conditions. The thermal profiles for each experimental single-track were simulated, enabling extraction of key solidification parameters, such as thermal gradient, and cooling rate, and analysis of their spatial-temporal evolution. Furthermore, the relationship between process parameters, thermal conditions, and cracking behavior were defined, identifying certain thermal profiles that reduce cracking severity.
Building on this foundation, the study progresses to single-layer deposition involving multiple tracks extending the single-track findings. Two distinct single-track process parameter recipes were selected to assess heat accumulation effects originating from the lateral stacking of tracks. The lateral arrangement was modulated using independent variables: line offset and track number. The impacts of these variables on deposited layer quality, multi-track microstructure, and cracking behaviour were assessed. The different single-track parameter recipes indicated contrasting multi-track deposition nature as reflected by surface topography analysis and subsequent assessment of the multi-track cross section. Moreover, a distinct spatial pattern of defects was observed. While these established relationships and findings point toward optimal multi-track conditions, further refinement and follow-up strategies are necessary to fully optimize the process