CPU-memory bottleneck is a widely recognized problem. It is known that majority of high performance computing (HPC) database systems are configured with large memories and dedicated to process specific workloads like weather prediction, molecular dynamic simulations etc. My research on optimal address mapping improves the memory performance by increasing the channel and bank level parallelism. In an another research direction, I proposed and evaluated adaptive page migration techniques that obviates the need for offline analysis of an application to determine page migration strategies. Furthermore, I explored different migration strategies like reverse migration, sub page migration that I found to be beneficial depending on the application behavior. Ideally, page migration strategies redirect the demand memory traffic to faster memory to improve the memory performance. In my third contribution, I worked and evaluated a memory-side accelerator to assist the main computational core in locating the non-zero elements of a sparse matrix that are typically used in scientific, machine learning workloads on a low-power embedded system configuration. Thus my contributions narrow the speed-gap by improving the latency and/or bandwidth between CPU and memory.

Prompt and efficient intervention is vital in reducing casualty figures during epidemic outbreaks, disasters, sudden civil strife or terrorism attacks. This can only be achieved if there is a fit-for-purpose and location-specific emergency response plan in place, incorporating geographical, time and vehicular capacity constraints. In this research, a comprehensive emergency response model for situations of uncertainties (in locations' demand and available resources), typically obtainable in low-resource countries, is designed. It involves the development of algorithms for optimizing pre-and post-disaster activities. The studies result in the development of four models: (1) an adaptation of a machine learning clustering algorithm, for pre-positioning depots and emergency operation centers, which optimizes the placement of these depots, such that the largest geographical location is covered, and the maximum number of individuals reached, with minimal facility cost; (2) an optimization algorithm for routing relief distribution, using heterogenous fleets of vehicle, with considerations for uncertainties in humanitarian supplies; (3) a genetic algorithm-based route improvement model; and (4) a model for integrating possible new locations into the routing network, in real-time, using emergency severity ranking, with a high priority on the most-vulnerable population. The clustering approach to solving dept location-allocation problem produces a better time complexity, and the …

Intelligent transportation systems (ITS) are essential tools for traffic planning, analysis, and forecasting that can utilize the huge amount of traffic data available nowadays. In this work, we aggregated detailed traffic flow sensor data, Waze reports, OpenStreetMap (OSM) features, and weather data, from California Bay Area for 6 months. Using that data, we studied three novel ITS applications using convolutional neural networks (CNNs) and recurrent neural networks (RNNs). The first experiment is an analysis of the relation between roadway shapes and accident occurrence, where results show that the speed limit and number of lanes are significant predictors for major accidents on highways. The second experiment presents a novel method for forecasting congestion severity using crowdsourced data only (Waze, OSM, and weather), without the need for traffic sensor data. The third experiment studies the improvement of traffic flow forecasting using accidents, number of lanes, weather, and time-related features, where results show significant performance improvements when the additional features where used.

Hateful content and offensive language are commonplace on social media platforms. Many surveys prove that high percentages of social media users experience online harassment. Previous efforts have been made to detect and remove online hate content automatically. However, removing users' content restricts free speech. A complementary strategy to address hateful content that does not interfere with free speech is to counter the hate with new content to divert the discourse away from the hate. In this dissertation, we complement the lack of previous work on counterhate arguments by analyzing and detecting them. Firstly, we study the relationships between hateful tweets and replies. Specifically, we analyze their fine-grained relationships by indicating whether the reply counters the hate, provides a justification, attacks the author of the tweet, or adds additional hate. The most obvious finding is that most replies generally agree with the hateful tweets; only 20% of them counter the hate. Secondly, we focus on the hate directed toward individuals and detect authentic counterhate arguments from online articles. We propose a methodology that assures the authenticity of the argument and its specificity to the individual of interest. We show that finding arguments in online articles is an efficient alternative compared to …

The goal of this thesis is to support the automated identification of accessibility in user reviews or bug reports, to help technology professionals prioritize their handling, and, thus, to create more inclusive apps. Particularly, we propose a model that takes as input accessibility user reviews or bug reports and learns their keyword-based features to make a classification decision, for a given review, on whether it is about accessibility or not. Our empirically driven study follows a mixture of qualitative and quantitative methods. We introduced models that can accurately identify accessibility reviews and bug reports and automate detecting them. Our models can automatically classify app reviews and bug reports as accessibility-related or not so developers can easily detect accessibility issues with their products and improve them to more accessible and inclusive apps utilizing the users' input. Our goal is to create a sustainable change by including a model in the developer's software maintenance pipeline and raising awareness of existing errors that hinder the accessibility of mobile apps, which is a pressing need. In light of our findings from the Blackboard case study, Blackboard and the course material are not easily accessible to deaf students and hard of hearing. Thus, deaf students …

The original FlexiChain and its descendants are a revolutionary distributed ledger technology (DLT) for cyber-physical systems (CPS) and their embedded systems (ES). FlexiChain, a DLT implementation, uses cryptography, distributed ledgers, peer-to-peer communications, scalable networks, and consensus. FlexiChain facilitates data structure agreements. This thesis offers a Block Directed Acyclic Graph (BDAG) architecture to link blocks to their forerunners to speed up validation. These data blocks are securely linked. This dissertation introduces Proof of Rapid Authentication, a novel consensus algorithm. This innovative method uses a distributed file to safely store a unique identifier (UID) based on node attributes to verify two blocks faster. This study also addresses CPS hardware security. A system of interconnected, user-unique identifiers allows each block's history to be monitored. This maintains each transaction and the validators who checked the block to ensure trustworthiness and honesty. We constructed a digital version that stays in sync with the distributed ledger as all nodes are linked by a NodeChain. The ledger is distributed without compromising node autonomy. Moreover, FlexiChain Layer 0 distributed ledger is also introduced and can connect and validate Layer 1 blockchains. This project produced a DAG-based blockchain integration platform with hardware security. The results illustrate a practical technique …

In this disseration we propose semi-automated top-down policy engineering approaches for attribute-based access control (ABAC) development. Further, we propose a hybrid ABAC policy engineering approach to combine the benefits and address the shortcomings of both top-down and bottom-up approaches. In particular, we propose three frameworks: (i) ABAC attributes extraction, (ii) ABAC constraints extraction, and (iii) hybrid ABAC policy engineering. Attributes extraction framework comprises of five modules that operate together to extract attributes values from natural language access control policies (NLACPs); map the extracted values to attribute keys; and assign each key-value pair to an appropriate entity. For ABAC constraints extraction framework, we design a two-phase process to extract ABAC constraints from NLACPs. The process begins with the identification phase which focuses on identifying the right boundary of constraint expressions. Next is the normalization phase, that aims at extracting the actual elements that pose a constraint. On the other hand, our hybrid ABAC policy engineering framework consists of 5 modules. This framework combines top-down and bottom-up policy engineering techniques to overcome the shortcomings of both approaches and to generate policies that are more intuitive and relevant to actual organization policies. With this, we believe that our work takes essential steps towards …

In this dissertation, I investigate the existing smart contract problems that limit cognitive abilities. I use Taylor's serious expansion, polynomial equation, and fraction-based computations to overcome the limitations of calculations in smart contracts. To prove the hypothesis, I use these mathematical models to compute complex operations of naive Bayes, linear regression, decision trees, and neural network algorithms on Ethereum public test networks. The smart contracts achieve 95\% prediction accuracy compared to traditional programming language models, proving the soundness of the numerical derivations. Many non-real-time applications can use our solution for trusted and secure prediction services.

In the first part of the dissertation, we provide an extensive overview of sub-6 GHz wireless access technology known as massive multiple-input multiple-output (MIMO) systems, highlighting its benefits, deployment challenges, and the key enabling technologies envisaged for 5G networks. We investigate the fundamental issues that degrade the performance of massive MIMO systems such as pilot contamination, precoding, user scheduling, and signal detection. In the second part, we optimize the performance of the massive MIMO system by proposing several algorithms, system designs, and hardware architectures. To mitigate the effect of pilot contamination, we propose a pilot reuse factor scheme based on the user environment and the number of active users. The results through simulations show that the proposed scheme ensures the system always operates at maximal spectral efficiency and achieves higher throughput. To address the user scheduling problem, we propose two user scheduling algorithms bases upon the measured channel gain. The simulation results show that our proposed user scheduling algorithms achieve better error performance, improve sum capacity and throughput, and guarantee fairness among the users. To address the uplink signal detection challenge in the massive MIMO systems, we propose four algorithms and their system designs. We show through simulations that the …

This dissertation first conducts a study on raw-data level cooperative perception for enhancing the detection ability of self-driving systems for connected autonomous vehicles (CAVs). A LiDAR (Light Detection and Ranging sensor) point cloud-based 3D object detection method is deployed to enhance detection performance by expanding the effective sensing area, capturing critical information in multiple scenarios and improving detection accuracy. In addition, a point cloud feature based cooperative perception framework is proposed on edge computing system for CAVs. This dissertation also uses the features' intrinsically small size to achieve real-time edge computing, without running the risk of congesting the network. In order to distinguish small sized objects such as pedestrian and cyclist in 3D data, an end-to-end multi-sensor fusion model is developed to implement 3D object detection from multi-sensor data. Experiments show that by solving multiple perception on camera and LiDAR jointly, the detection model can leverage the advantages from high resolution image and physical world LiDAR mapping data, which leads the KITTI benchmark on 3D object detection. At last, an application of cooperative perception is deployed on edge to heal the live map for autonomous vehicles. Through 3D reconstruction and multi-sensor fusion detection, experiments on real-world dataset demonstrate that a …

Possession is an asymmetric semantic relation between two entities, where one entity (the possessee) belongs to the other entity (the possessor). Automatically extracting possessions are useful in identifying skills, recommender systems and in natural language understanding. Possessions can be found in different communication modalities including text, images, videos, and audios. In this dissertation, I elaborate on the techniques I used to extract possessions. I begin with extracting possessions at the sentence level including the type and temporal anchors. Then, I extract the duration of possession and co-possessions (if multiple possessors possess the same entity). Next, I extract possessions from an entire Wikipedia article capturing the change of possessors over time. I extract possessions from social media including both text and images. Finally, I also present dense annotations generating possession timelines. I present separate datasets, detailed corpus analysis, and machine learning models for each task described above.

The vehicular community is moving towards a whole new paradigm with the advancement of new technology. Vehicular communication not only supports safety services but also provides non-safety services like navigation support, toll collection, web browsing, media streaming, etc. The existing communication frameworks like Dedicated Short Range Communication (DSRC) and Cellular V2X (C-V2X) might not meet the required capacity in the coming days. So, the vehicular community needs to adopt new technologies and upgrade the existing communication frameworks so that it can fulfill the desired expectations. Therefore, an increment in reliability and data rate is required. Multiple Input Multiple Output (MIMO), 5G New Radio, Low Density Parity Check (LDPC) Code, and Massive MIMO signal detection and equalization algorithms are the latest addition to the 5G wireless communication domain. These technologies have the potential to make the existing V2X communication framework more robust. As a result, more reliability and throughput can be achieved. This work demonstrates these technologies' compatibility and positive impact on existing V2X communication standard.

In this work, we leverage the recent developments in computer science to address several of the challenges in current literature-based discovery (LBD) solutions. First, LBD solutions cannot use semantics or are too computational complex. To solve the problems we propose a generative model OverlapLDA based on topic modeling, which has been shown both effective and efficient in extracting semantics from a corpus. We also introduce an inference method of OverlapLDA. We conduct extensive experiments to show the effectiveness and efficiency of OverlapLDA in LBD. Second, we expand LBD to a more complex and realistic setting. The settings are that there can be more than one concept connecting the input concepts, and the connectivity pattern between concepts can also be more complex than a chain. Current LBD solutions can hardly complete the LBD task in the new setting. We simplify the hypotheses as concept sets and propose LBDSetNet based on graph neural networks to solve this problem. We also introduce different training schemes based on self-supervised learning to train LBDSetNet without relying on comprehensive labeled hypotheses that are extremely costly to get. Our comprehensive experiments show that LBDSetNet outperforms strong baselines on simple hypotheses and addresses complex hypotheses.

The main research objective of this dissertation is to understand the sensing and communication challenges to achieving cooperative perception among autonomous vehicles, and then, using the insights gained, guide the design of the suitable format of data to be exchanged, reliable and efficient data fusion algorithms on vehicles. By understanding what and how data are exchanged among autonomous vehicles, from a machine learning perspective, it is possible to realize precise cooperative perception on autonomous vehicles, enabling massive amounts of sensor information to be shared amongst vehicles. I first discuss the trustworthy perception information sharing on connected and autonomous vehicles. Then how to achieve effective cooperative perception on autonomous vehicles via exchanging feature maps among vehicles is discussed in the following. In the last methodology part, I propose a set of mechanisms to improve the solution proposed before, i.e., reducing the amount of data transmitted in the network to achieve an efficient cooperative perception. The effectiveness and efficiency of our mechanism is analyzed and discussed.

Machine learning algorithms based on neural networks have achieved remarkable results and are being extensively used in different domains. However, the machine learning algorithms requires access to raw data which is often privacy sensitive. To address this issue, we develop new techniques to provide solutions for running deep neural networks over encrypted data. In this paper, we develop new techniques to adopt deep neural networks within the practical limitation of current homomorphic encryption schemes. We focus on training and classification of the well-known neural networks and convolutional neural networks. First, we design methods for approximation of the activation functions commonly used in CNNs (i.e. ReLU, Sigmoid, and Tanh) with low degree polynomials which is essential for efficient homomorphic encryption schemes. Then, we train neural networks with the approximation polynomials instead of original activation functions and analyze the performance of the models. Finally, we implement neural networks and convolutional neural networks over encrypted data and measure performance of the models.

Autonomous vehicles have made milestone strides within the past decade. Advances up the autonomy ladder have come lock-step with the advances in machine learning, namely deep-learning algorithms and huge, open training sets. And while advances in CPUs have slowed, GPUs have edged into the previous decade's TOP 500 supercomputer territory. This new class of GPUs include novel deep-learning hardware that has essentially side-stepped Moore's law, outpacing the doubling observation by a factor of ten. While GPUs have make record progress, networks do not follow Moore's law and are restricted by several bottlenecks, from protocol-based latency lower bounds to the very laws of physics. In a way, the bottlenecks that plague modern networks gave rise to Edge computing, a key component of the Connected Autonomous Vehicle system, as the need for low-latency in some domains eclipsed the need for massive processing farms. The Connected Autonomous Vehicle ecosystem is one of the most complicated environments in all of computing. Not only is the hardware scaled all the way from 16 and 32-bit microcontrollers, to multi-CPU Edge nodes, and multi-GPU Cloud servers, but the networking also encompasses the gamut of modern communication transports. I propose a framework for negotiating, encapsulating and transferring data …

In this dissertation, I start with an analysis of negation in eleven benchmark corpora covering six Natural Language Understanding (NLU) tasks. With a thorough investigation, I first show that (a) these benchmarks contain fewer negations compared to general-purpose English and (b) the few negations they contain are often unimportant. Further, my empirical studies demonstrate that state-of-the-art transformers trained using these corpora obtain substantially worse results with the instances that contain negation, especially if the negations are important. Second, I investigate whether translating negation is also an issue for modern machine translation (MT) systems. My studies find that indeed the presence of negation can significantly impact translation quality, in some cases resulting in reductions of over 60%. In light of these findings, I investigate strategies to better understand the semantics of negation. I start with identifying the focus of negation. I develop a neural model that takes into account the scope of negation, context from neighboring sentences, or both. My best proposed system obtains an accuracy improvement of 7.4% over prior work. Further, I analyze the main error categories of the systems through a detailed error analysis. Next, I explore more practical ways to understand the semantics of negation. I consider …

The novelty of this dissertation is the optimization and acceleration of deep neural networks aimed at real-time predictions with minimal energy consumption. It consists of cross-layer optimization, output directed dynamic quantization, and opportunistic near-data computation for deep neural network acceleration. On two datasets (CIFAR-10 and CIFAR-100), the proposed deep neural network optimization and acceleration frameworks are tested using a variety of Convolutional neural networks (e.g., LeNet-5, VGG-16, GoogLeNet, DenseNet, ResNet). Experimental results are promising when compared to other state-of-the-art deep neural network acceleration efforts in the literature.

With increasing data and technology, multiple types of data from the same set of nodes have been generated. Since each data modality contains a unique aspect of the underlying mechanisms, multiple datatypes are integrated. In addition to multiple datatypes, networks are important to store information representing associations between entities such as genes of a protein-protein interaction network and authors of a citation network. Recently, some advanced approaches to graph-structured data leverage node associations and features simultaneously, called Graph Neural Network (GNN), but they have limitations for integrative approaches. The overall aim of this dissertation is to integrate multiple data modalities on graph-structured data to infer some context-specific gene regulation and predict outcomes of interest. To this end, first, we introduce a computational tool named CRINET to infer genome-wide competing endogenous RNA (ceRNA) networks. By integrating multiple data properly, we had a better understanding of gene regulatory circuitry addressing important drawbacks pertaining to ceRNA regulation. We tested CRINET on breast cancer data and found that ceRNA interactions and groups were significantly enriched in the cancer-related genes and processes. CRINET-inferred ceRNA groups supported the studies claiming the relation between immunotherapy and cancer. Second, we present SUPREME, a node classification framework, by comprehensively …

This dissertation introduces a hybrid strategy for automated testing of Android applications that combines reinforcement learning and test suite prioritization. These approaches aim to improve the effectiveness of the testing process by employing reinforcement learning algorithms, namely Q-learning and SARSA (State-Action-Reward-State-Action), for automated test case generation. The studies provide compelling evidence that reinforcement learning techniques hold great potential in generating test cases that consistently achieve high code coverage; however, the generated test cases may not always be in the optimal order. In this study, novel test case prioritization methods are developed, leveraging pairwise event interactions coverage, application state coverage, and application activity coverage, so as to optimize the rates of code coverage specifically for SARSA-generated test cases. Additionally, test suite prioritization techniques are introduced based on UI element coverage, test case cost, and test case complexity to further enhance the ordering of SARSA-generated test cases. Empirical investigations demonstrate that applying the proposed test suite prioritization techniques to the test suites generated by the reinforcement learning algorithm SARSA improved the rates of code coverage over original orderings and random orderings of test cases.

Recently, disaster scene description and indexing challenges attract the attention of researchers. In this dissertation, we solve a disaster-related multi-labeling task using a newly developed Low Altitude Disaster Imagery dataset. In the first task, we realize video content by selecting a set of summary key frames to represent the video sequence. Through inter-frame differences, the key frames are generated. The key frame extraction of disaster-related video clips is a powerful tool that can efficiently convert video data into image-level data, reduce the requirements for the extraction environment and improve the applicable environment. In the second, we propose a novel application of using deep learning methods on low altitude disaster video feature recognition. Supervised learning-based deep-learning approaches are effective in disaster-related features recognition via foreground object detection and background classification. Performed dataset validation, our model generalized well and improved performance by optimizing the YOLOv3 model and combining it with Resnet50. The comprehensive models showed more efficient and effective than those in prior published works. In the third task, we optimize the whole scene labeling classification by pruning the lightweight model MobileNetV3, which shows superior generalizability and can disaster features recognition from a disaster-related dataset be accomplished efficiently to assist disaster recovery.

NAND flash-based solid state drives (SSDs) have been widely adopted in data centers and high performance computing (HPC) systems due to their better performance compared with hard disk drives. However, little is known about the reliability characteristics of SSDs in production systems. Existing works that study the statistical distributions of SSD failures in the field lack insights into distinct characteristics of SSDs. In this dissertation, I explore the SSD-specific SMART (Self-Monitoring, Analysis, and Reporting Technology) attributes and conduct in-depth analysis of SSD reliability in a production environment with a focus on the unique error types and health dynamics. QLC SSD delivers better performance in a cost-effective way. I study QLC SSDs in terms of their architecture and performance. In addition, I apply thermal stress tests to QLC SSDs and quantify their performance degradation processes. Various types of big data and machine learning workloads have been executed on SSDs under varying temperatures. The SSD throughput and application performance are analyzed and characterized.

With the advent of autonomous systems and deep learning systems, safety pertaining to these systems has become a major concern. The existing failure analysis techniques are not enough to thoroughly analyze the safety in these systems. Moreover, because these systems are created to operate in various conditions, they are susceptible to unknown safety issues. Hence, we need mechanisms which can take into account the complexity of operational design domains, identify safety issues other than failures, and expose unknown safety issues. Moreover, existing safety analysis approaches require a lot of effort and time for analysis and do not consider machine learning (ML) safety. To address these limitations, in this dissertation, we discuss an artificial-intelligence driven model-based methodology that aids in identifying unknown safety issues and analyzing ML safety. Our methodology consists of 4 major tasks: 1) automated model generation, 2) automated analysis of component state transition model specification, 3) undesired states analysis, and 4) causal factor analysis. In our methodology we identify unknown safety issues by finding undesired combinations of components' states and environmental entities' states as well as causes resulting in these undesired combinations. In our methodology, we refer to the behaviors that occur because of undesired combinations as off-nominal …

Non-rigid point set registration of significantly uneven deformations is a challenging problem for many applications such as pose estimation, three-dimensional object reconstruction, human movement tracking. In this dissertation, we present a novel probabilistic non-rigid registration method to align point sets with significantly uneven deformations by enforcing constraints from corresponding key points and preserving local neighborhood structures. The registration method is treated as a density estimation problem. Incorporating correspondence among key points regulates the optimization process for large, uneven deformations. In addition, by leveraging neighborhood embedding using Stochastic Neighbor Embedding (SNE) as well as an alternative means based on Locally Linear Embedding (LLE), our method penalizes the incoherent transformation and hence preserves the local structure of point sets. Also, our method detects key points in the point sets based on geodesic distance. Correspondences are established using a new cluster-based, region-aware feature descriptor. This feature descriptor encodes the association of a cluster to the left-right (symmetry) or upper-lower regions of the point sets. We conducted comparison studies using public point sets and our Human point sets. Our experimental results demonstrate that our proposed method successfully reduced the registration error by at least 42.2% in contrast to the state-of-the-art method. Especially, our method …