Open Access. Powered by Scholars. Published by Universities.®
Physical Sciences and Mathematics Commons™
Open Access. Powered by Scholars. Published by Universities.®
Articles 1 - 30 of 39
Full-Text Articles in Physical Sciences and Mathematics
The Dollar General: Continuous Custom Gesture Recognition Techniques At Everyday Low Prices, Eugene Taranta
The Dollar General: Continuous Custom Gesture Recognition Techniques At Everyday Low Prices, Eugene Taranta
Electronic Theses and Dissertations, 2020-
Humans use gestures to emphasize ideas and disseminate information. Their importance is apparent in how we continuously augment social interactions with motion—gesticulating in harmony with nearly every utterance to ensure observers understand that which we wish to communicate, and their relevance has not escaped the HCI community's attention. For almost as long as computers have been able to sample human motion at the user interface boundary, software systems have been made to understand gestures as command metaphors. Customization, in particular, has great potential to improve user experience, whereby users map specific gestures to specific software functions. However, custom gesture recognition …
Algorithms And Applications Of Novel Capsule Networks, Rodney Lalonde
Algorithms And Applications Of Novel Capsule Networks, Rodney Lalonde
Electronic Theses and Dissertations, 2020-
Convolutional neural networks, despite their profound impact in countless domains, suffer from significant shortcomings. Linearly-combined scalar feature representations and max pooling operations lead to spatial ambiguities and a lack of robustness to pose variations. Capsule networks can potentially alleviate these issues by storing and routing the pose information of extracted features through their architectures, seeking agreement between the lower-level predictions of higher-level poses at each layer. In this dissertation, we make several key contributions to advance the algorithms of capsule networks in segmentation and classification applications. We create the first ever capsule-based segmentation network in the literature, SegCaps, by introducing …
Efficient String Algorithms With Applications In Bioinformatics, Sahar Hooshmand
Efficient String Algorithms With Applications In Bioinformatics, Sahar Hooshmand
Electronic Theses and Dissertations, 2020-
The work presented in this dissertation deals with establishing efficient methods for solving some algorithmic problems, which have applications to Bioinformatics. After a short introduction in Chapter 1, an algorithm for genome mappability problem is presented in Chapter 2. Genome mappability is a measure for the approximate repeat structure of the genome with respect to substrings of specific length and a tolerance to define the number of mismatches. The similarity between reads is measured by using the Hamming distance function. Genome mappability is computed for each position in the string and has several applications in designing high-throughput short-read sequencing experiments. …
The Effects Of Gesture Presentation In Video Games, Jack Oakley
The Effects Of Gesture Presentation In Video Games, Jack Oakley
Electronic Theses and Dissertations, 2020-
As everyday and commonplace technology continues to move toward touch devices and virtual reality devices, more and more video games are using gestures as forms of gameplay. While there is much research focused on gestures as user interface navigation methods, we wanted to look into how gestures affect gameplay when used as a gameplay mechanic. In particular, we set out to determine how different ways of presenting gestures might affect the game's difficulty and flow. We designed two versions of a zombie game where the zombies are killed by drawing gestures. The first version of the game is a touchscreen-based …
Action Recognition In Still Images: Confluence Of Multilinear Methods And Deep Learning, Marjaneh Safaei
Action Recognition In Still Images: Confluence Of Multilinear Methods And Deep Learning, Marjaneh Safaei
Electronic Theses and Dissertations, 2020-
Motion is a missing information in an image, however, it is a valuable cue for action recognition. Thus, lack of motion information in a single image makes action recognition for still images inherently a very challenging problem in computer vision. In this dissertation, we show that both spatial and temporal patterns provide crucial information for recognizing human actions. Therefore, action recognition depends not only on the spatially-salient pixels, but also on the temporal patterns of those pixels. To address the challenge caused by the absence of temporal information in a single image, we introduce five effective action classification methodologies along …
Reconstruction Of Bacterial Strain Genomes From Shotgun Metagenomic Reads, Xin Li
Reconstruction Of Bacterial Strain Genomes From Shotgun Metagenomic Reads, Xin Li
Electronic Theses and Dissertations, 2020-
It is necessary to study bacterial strains in environmental samples. The environmental samples are mixed DNA samples collected from the ocean, soil, lake, human body sites, etc. In a natural environment, they provide us new insights into the diversity of our earth. As for bacterial strains on or inside human bodies, to select the proper treatment for diseases caused by bacterial strains, it is critical to identify the corresponding strains and reconstruct their genomes. However, it is a challenge to do so with the DNA from a large number of unknown microbial species mixed together in an environmental sample. The …
Computational Methods For Discovery And Analysis Of Rna Structural Motifs, Shahidul Islam
Computational Methods For Discovery And Analysis Of Rna Structural Motifs, Shahidul Islam
Electronic Theses and Dissertations, 2020-
Understanding the 3D structural properties of RNAs will play a critical role in identifying their functional characteristics and designing new RNAs for RNA-based therapeutics and nanotechnology. In an attempt to achieve a better insight into RNAs, biochemical experiments have been conducted to produce data with positional details of atoms in RNA structures. This data has created opportunities for applying computational analysis to solve various biological problems. In this dissertation, we have addressed annotation issues of base-pairing interactions in the low-resolution structure data and presented new methods to analyze RNA structural motifs. Annotating base-pairing interactions is one of the critical steps …
Algorithms For Inferring Multiple Microbial Networks, Sahar Tavakoli
Algorithms For Inferring Multiple Microbial Networks, Sahar Tavakoli
Electronic Theses and Dissertations, 2020-
The interactions among the constituent members of a microbial community play a major role in determining the overall behavior of the community and the abundance levels of its members. These interactions can be modeled using a network whose nodes represent microbial taxa and edges represent pairwise interactions. A microbial network is a weighted graph that is constructed from a sample-taxa count matrix and can be used to model co-occurrences and/or interactions of the constituent members of a microbial community. The nodes in this graph represent microbial taxa and the edges represent pairwise associations amongst these taxa. A microbial network is …
Equivariance And Invariance For Robust Unsupervised And Semi-Supervised Learning, Liheng Zhang
Equivariance And Invariance For Robust Unsupervised And Semi-Supervised Learning, Liheng Zhang
Electronic Theses and Dissertations, 2020-
Although there is a great success of applying deep learning on a wide variety of tasks, it heavily relies on a large amount of labeled training data, which could be hard to obtain in many real scenarios. To address this problem, unsupervised and semi-supervised learning emerge to take advantage of the plenty of cheap unlabeled data to improve the model generalization. In this dissertation, we claim that equivariant and invariance are two critical criteria to approach robust unsupervised and semi-supervised learning. The idea is as follows: the features of a robust model ought to be sufficiently informative and equivariant to …
3d Localization Of Defects In Facility Inspections, Nicholas Califano
3d Localization Of Defects In Facility Inspections, Nicholas Califano
Electronic Theses and Dissertations, 2020-
Wind tunnels are crucial facilities that support the aerospace industry. However, these facilities are large, complex, and pose unique maintenance and inspection requirements. Manual inspections to identify defects such as cracks, missing fasteners, leaks, and foreign objects are important but labor and schedule intensive. The goal of this thesis is to utilize small Unmanned Aircraft Systems with onboard cameras and computer vision-based analysis to automate the inspection of the interior and exterior of NASA's critical wind tunnel facilities. Missing fasteners are detected as the defect class, and existing fasteners are detected to provide potential future missing fastener sites for preventative …
Novel Computational Approaches For Multidimensional Brain Image Analysis, Harish Raviprakash
Novel Computational Approaches For Multidimensional Brain Image Analysis, Harish Raviprakash
Electronic Theses and Dissertations, 2020-
The overall goal of this dissertation is focused on addressing challenging problems in 1D, 2D/3D and 4D neuroimaging by developing novel algorithms that combine signal processing and machine learning techniques. One of these challenging tasks is the accurate localization of the eloquent language cortex in brain resection pre-surgery patients. This is especially important since inaccurate localization can lead to diminshed functionalities and thus, a poor quality of life for the patient. The first part of this dissertation addresses this problem in the case of drug-resistant epileptic patients. We propose a novel machine learning based algorithm to establish an alternate electrical …
Analyzing User Behavior In Collaborative Environments, Samaneh Saadat
Analyzing User Behavior In Collaborative Environments, Samaneh Saadat
Electronic Theses and Dissertations, 2020-
Discrete sequences are the building blocks for many real-world problems in domains including genomics, e-commerce, and social sciences. While there are machine learning methods to classify and cluster sequences, they fail to explain what makes groups of sequences distinguishable. Although in some cases having a black box model is sufficient, there is a need for increased explainability in research areas focused on human behaviors. For example, psychologists are less interested in having a model that predicts human behavior with high accuracy and more concerned with identifying differences between actions that lead to divergent human behavior. This dissertation presents techniques for …
Open-Ended Search Through Minimal Criterion Coevolution, Jonathan Brant
Open-Ended Search Through Minimal Criterion Coevolution, Jonathan Brant
Electronic Theses and Dissertations, 2020-
Search processes guided by objectives are ubiquitous in machine learning. They iteratively reward artifacts based on their proximity to an optimization target, and terminate upon solution space convergence. Some recent studies take a different approach, capitalizing on the disconnect between mainstream methods in artificial intelligence and the field's biological inspirations. Natural evolution has an unparalleled propensity for generating well-adapted artifacts, but these artifacts are decidedly non-convergent. This new class of non-objective algorithms induce a divergent search by rewarding solutions according to their novelty with respect to prior discoveries. While the diversity of resulting innovations exhibit marked parallels to natural evolution, …
Video Content Understanding Using Text, Amir Mazaheri
Video Content Understanding Using Text, Amir Mazaheri
Electronic Theses and Dissertations, 2020-
The rise of the social media and video streaming industry provided us a plethora of videos and their corresponding descriptive information in the form of concepts (words) and textual video captions. Due to the mass amount of available videos and the textual data, today is the best time ever to study the Computer Vision and Machine Learning problems related to videos and text. In this dissertation, we tackle multiple problems associated with the joint understanding of videos and text. We first address the task of multi-concept video retrieval, where the input is a set of words as concepts, and the …
Endpoints And Interdependencies In Internet Of Things Residual Artifacts: Measurements, Analyses, And Insights Into Defenses, Jinchun Choi
Endpoints And Interdependencies In Internet Of Things Residual Artifacts: Measurements, Analyses, And Insights Into Defenses, Jinchun Choi
Electronic Theses and Dissertations, 2020-
The usage of Internet of Things (IoT) devices is growing fast. Moreover, the lack of security measures among the IoT devices and their persistent online connection give adversaries an opportunity to exploit them for multiple types of attacks, such as distributed denial-of-service (DDoS). To understand the risks of IoT devices, we analyze IoT malware from an endpoint standpoint. We investigate the relationship between endpoints infected and attacked by IoT malware, and gain insights into the underlying dynamics in the malware ecosystem. We observe the affinities and different patterns among endpoints. Towards this, we reverse-engineer 2,423 IoT malware samples and extract …
Improving The Security Of Critical Infrastructure: Metrics, Measurements, And Analysis, Jeman Park
Improving The Security Of Critical Infrastructure: Metrics, Measurements, And Analysis, Jeman Park
Electronic Theses and Dissertations, 2020-
In this work, we propose three important contributions needed in the process of improving the security of the critical infrastructure: metrics, measurement, and analysis. To improve security, metrics are key to ensuring the accuracy of the assessment and evaluation. Measurements are the core of the process of identifying the causality and effectiveness of various behaviors, and accurate measurement with the right assumptions is a cornerstone for accurate analysis. Finally, contextualized analysis essential for understanding measurements. Different results can be derived for the same data according to the analysis method, and it can serve as a basis for understanding and improving …
Separating Content Selection From Surface Realization In Neural Text Summarization, Logan Lebanoff
Separating Content Selection From Surface Realization In Neural Text Summarization, Logan Lebanoff
Electronic Theses and Dissertations, 2020-
Text summarization is a rapidly growing field with many new innovations. End-to-end models using the sequence-to-sequence architecture achieve high scores according to automatic metrics on standard datasets. However, they frequently generate summaries that are factually inconsistent with the original article -- a vital problem to be solved before the summaries can be used in real-world applications. In addition, they are not generalizable to new domains, especially those with few training examples. In this dissertation, we propose to explicitly separate the two steps of content selection and surface realization in summarization. Content selection is the process of choosing important words/phrases/sentences from …
Improving Security Of Crypto Wallets In Blockchain Technologies, Hossein Rezaeighaleh
Improving Security Of Crypto Wallets In Blockchain Technologies, Hossein Rezaeighaleh
Electronic Theses and Dissertations, 2020-
A big challenge in blockchain and cryptocurrency is securing the private key from potential hackers. Nobody can rollback a transaction made with a stolen key once the network confirms it. The technical solution to protect private keys is the cryptocurrency wallet, software, hardware, or a combination to manage the keys. In this dissertation, we try to investigate the significant challenges in existing cryptocurrency wallets and propose innovative solutions. Firstly, almost all cryptocurrency wallets suffer from the lack of a secure and convenient backup and recovery process. We offer a new cryptographic scheme to securely back up a hardware wallet relying …
Towards Large-Scale And Robust Code Authorship Identification With Deep Feature Learning, Mohammed Abuhamad
Towards Large-Scale And Robust Code Authorship Identification With Deep Feature Learning, Mohammed Abuhamad
Electronic Theses and Dissertations, 2020-
Successful software authorship identification has both software forensics applications and privacy implications. However, the process requires an efficient extraction of quality authorship attributes. The extraction of such attributes is very challenging due to several factors such as the variety of software formats, number of available samples, and possible obfuscation or adversarial manipulation. We focus on software authorship identification from three central perspectives: large-scale single-authored software, real-world multi-authored software, and the robustness assessment of code authorship identification methods against adversarial attacks. First, we propose DL-CAIS, a deep Learning-based approach for software authorship attribution, that facilitates large-scale, format-independent, language-oblivious, and obfuscation-resilient software …
Deep Recurrent Networks For Gesture Recognition And Synthesis, Mehran Maghoumi
Deep Recurrent Networks For Gesture Recognition And Synthesis, Mehran Maghoumi
Electronic Theses and Dissertations, 2020-
It is hard to overstate the importance of gesture-based interfaces in many applications nowadays. The adoption of such interfaces stems from the opportunities they create for incorporating natural and fluid user interactions. This highlights the importance of having gesture recognizers that are not only accurate but also easy to adopt. The ever-growing popularity of machine learning has prompted many application developers to integrate automatic methods of recognition into their products. On the one hand, deep learning often tops the list of the most powerful and robust recognizers. These methods have been consistently shown to outperform all other machine learning methods …
Interdisciplinary Cybersecurity For Resilient Cyberdefense, Rachid Ait Maalem Lahcen
Interdisciplinary Cybersecurity For Resilient Cyberdefense, Rachid Ait Maalem Lahcen
Electronic Theses and Dissertations, 2020-
Cybersecurity's role is to protect confidentiality, integrity, and availability of enterprise assets. Confidentiality secures data from theft, integrity mitigates modification of data in a malicious way, and availability assures continuation of systems' access and services. However, achieving these goals is difficult due to the mushrooming of various cyber attackers that come from individuals or state actors with motives ranging from ideological, financial, state-sponsored espionage, revenge, or simple curiosity and boredom. The difficulty also lies in the complexity of the cyber layers that are not well studied. Layers that interconnect and require effective communication and collaboration. This effectiveness is still lacking …
Interactions Between Humans, Virtual Agent Characters And Virtual Avatars, Tamara Griffith
Interactions Between Humans, Virtual Agent Characters And Virtual Avatars, Tamara Griffith
Electronic Theses and Dissertations, 2020-
Simulations allow people to experience events as if they were happening in the real world in a way that is safer and less expensive than live training. Despite improvements in realism in simulated environments, one area that still presents a challenge is interpersonal interactions. The subtleties of what makes an interaction rich are difficult to define. We may never fully understand the complexity of human interchanges, however there is value in building on existing research into how individuals react to virtual characters to inform future investments. Virtual characters can either be automated through computational processes, referred to as agents, or …
Learning Transferable Representations For Visual Recognition, Yang Zhang
Learning Transferable Representations For Visual Recognition, Yang Zhang
Electronic Theses and Dissertations, 2020-
In the last half-decade, a new renaissance of machine learning originates from the applications of convolutional neural networks to visual recognition tasks. It is believed that a combination of big curated data and novel deep learning techniques can lead to unprecedented results. However, the increasingly large training data is still a drop in the ocean compared with scenarios in the wild. In this literature, we focus on learning transferable representation in the neural networks to ensure the models stay robust, even given different data distributions. We present three exemplar topics in three chapters, respectively: zero-shot learning, domain adaptation, and generalizable …
High Performance And Secure Execution Environments For Emerging Architectures, Mazen Alwadi
High Performance And Secure Execution Environments For Emerging Architectures, Mazen Alwadi
Electronic Theses and Dissertations, 2020-
Energy-efficiency and performance have been the driving forces of system architectures and designers in the last century. Given the diversity of workloads and the significant performance and power improvements when running workloads on customized processing elements, system vendors are drifting towards new system architectures (e.g., FAM or HMM). Such architectures are being developed with the purpose of improving the system's performance, allow easier data sharing, and reduce the overall power consumption. Additionally, current computing systems suffer from a very wide attack surface, mainly due to the fact that such systems comprise of tens to hundreds of sub-systems that could be …
Design Of Ternary Operations Utilizing Flow-Based Computing, James Pyrich
Design Of Ternary Operations Utilizing Flow-Based Computing, James Pyrich
Electronic Theses and Dissertations, 2020-
The development of algorithms and circuit designs that exploit devices that have the ability to persist multiple values will lead to alternative technologies to overcome the issues caused by the end of Dennard scaling and slowing of Moore's Law. Flow-based designs have been used to develop binary adders and multipliers. Data stored on non-volatile memristors are used to direct the flow of current through nanowires arranged in a crossbar. The algorithmic design of the flow-based crossbar is fast, compact, and efficient. In this paper, we seek to automate the discovery of flow-based designs of ternary circuits utilizing memristive crossbars.
Towards Robust Artificial Intelligence Systems, Sunny Raj
Towards Robust Artificial Intelligence Systems, Sunny Raj
Electronic Theses and Dissertations, 2020-
Adoption of deep neural networks (DNNs) into safety-critical and high-assurance systems has been hindered by the inability of DNNs to handle adversarial and out-of-distribution input. State-of-the-art DNNs misclassify adversarial input and give high confidence output for out-of-distribution input. We attempt to solve this problem by employing two approaches, first, by detecting adversarial input and, second, by developing a confidence metric that can indicate when a DNN system has reached its limits and is not performing to the desired specifications. The effectiveness of our method at detecting adversarial input is demonstrated against the popular DeepFool adversarial image generation method. On a …
The Susceptibility Of Deep Neural Networks To Natural Perturbations, Mesut Ozdag
The Susceptibility Of Deep Neural Networks To Natural Perturbations, Mesut Ozdag
Electronic Theses and Dissertations, 2020-
Deep learning systems have achieved great success in various types of applications in recent years. They are increasingly being adopted for safety-critical tasks such as face recognition, surveillance systems, speech recognition, and autonomous driving. On the other hand, it has been found that deep neural networks (DNNs) can easily be fooled by adversarial input samples. These imperceptible perturbations on images can lead any machine learning system to misclassify the objects with high confidence. Furthermore, they can be almost indistinguishable to a human observer. These systems can also be exposed to adverse weather conditions such as fog, rain, and snow. This …
Explore And Design Novel Structures For More Efficient And Better Deep Convolutional Neural Networks, Min Wang
Explore And Design Novel Structures For More Efficient And Better Deep Convolutional Neural Networks, Min Wang
Electronic Theses and Dissertations, 2020-
Deep Convolutional Neural Networks have achieved remarkable performance on visual recognition problems, and have been extensively adopted in real-world applications, such as Apple's Face ID security system, autonomous driving cars, and automatic image tagging in online album services. One major concern in the development of CNNs is that their computational complexity grows along with the increase in their accuracy. Therefore, there is a continuous demand to find the right balance between accuracy and complexity in the design of CNN models. This dissertation focuses on designing various novel structures to enhance the performance of CNNs and their efficiency. Our efforts fall …
Decentralized Adaptable Task Allocation For Ongoing Tasks, Vera Kazakova
Decentralized Adaptable Task Allocation For Ongoing Tasks, Vera Kazakova
Electronic Theses and Dissertations, 2020-
This thesis extends an existing bio-inspired model for decentralized task allocation and benchmarks it against alternative approaches to assess robustness in dynamic conditions, applicability to domains with ongoing and hierarchical tasks, and scalability to large teams of agents. The work addresses decentralized task allocation of simple non-communicating agents in dynamic environments of multiple tasks. Multi-area patrolling is used as the sample domain: specific number of agents is required to successfully patrol each area on each timestep, indefinitely, until the system's security needs. Agents must individually decide whether to patrol and where (i.e., task availability is not limited by task demand …
Differential Recurrent Neural Networks For Human Activity Recognition, Naifan Zhuang
Differential Recurrent Neural Networks For Human Activity Recognition, Naifan Zhuang
Electronic Theses and Dissertations, 2020-
Human activity recognition has been an active research area in recent years. The difficulty of this problem lies in the complex dynamical motion patterns embedded through the sequential frames. The Long Short-Term Memory (LSTM) recurrent neural network is capable of processing complex sequential information since it utilizes special gating schemes for learning representations from long input sequences. It has the potential to model various time-series data, where the current hidden state has to be considered in the context of the past hidden states. Unfortunately, the conventional LSTMs do not consider the impact of spatio-temporal dynamics corresponding to the given salient …