Computer Sciences Commons^™

Multi-Task Zipping Via Layer-Wise Neuron Sharing, Xiaoxi He, Zimu Zhou, Lothar Thiele

Research Collection School Of Computing and Information Systems

Future mobile devices are anticipated to perceive, understand and react to the world on their own by running multiple correlated deep neural networks on-device. Yet the complexity of these neural networks needs to be trimmed down both within-model and cross-model to fit in mobile storage and memory. Previous studies squeeze the redundancy within a single model. In this work, we aim to reduce the redundancy across multiple models. We propose Multi-Task Zipping (MTZ), a framework to automatically merge correlated, pre-trained deep neural networks for cross-model compression. Central in MTZ is a layer-wise neuron sharing and incoming weight updating scheme that …

A Speech Recognition Noise Analysis Using A Biological Neural Network, Logan Matthew Walters

Student Theses and Dissertations

Typically, modern Automatic Speech Recognition (ASR) engines are developed using artificial neural networks (ANN) that utilize probabilistic and stochastic models for analyzing data and computing outputs. As such, limitations are seen with this structure of implementation due to the volume of data required for accuracy and scalability. Resultantly, different software solutions are being considered to improve the accuracy of ASR engines in these unique cases where commonly used models fall short. The presented research illustrates the potential of a biological neural network (BNN) if implemented as an ASR engine. Theoretical comparisons are made between the efficiency of a BNN and …

Change Descriptors For Determining Nodule Malignancy In Lung Ct Screening Images, Benjamin Geiger

USF Tampa Graduate Theses and Dissertations

Computed tomography (CT) imagery is an important weapon in the fight against lung cancer; various forms of lung cancer are routinely diagnosed from CT imagery. The growth of the suspect nodule is known to be a prognostic factor in the diagnosis of pulmonary cancer, but the change in other aspects of the nodule, such as its aspect ratio, density, spiculation, or other features usable for machine learning, may also provide prognostic information.

We hypothesized that adding combined feature information from multiple CT image sets separated in time could provide a more accurate determination of nodule malignancy. To this end, we …

User Attitudes About Duo Two-Factor Authentication At Byu, Jonathan Dutson

Undergraduate Honors Theses

Simple password-based authentication provides insufficient protection against increasingly common incidents of online identity theft and data loss. Although two-factor authentication (2FA) provides users with increased protection against attackers, users have mixed feelings about the usability of 2FA. We surveyed the students, faculty, and staff of Brigham Young University (BYU) to measure user sentiment about DUO Security, the 2FA system adopted by BYU in 2017. We find that most users consider DUO to be annoying, and about half of those surveyed expressed a preference for authentication without using a second-factor. About half of all participants reported at least one instance of …

College Of Engineering Senior Design Competition Fall 2018, University Of Nevada, Las Vegas

Fred and Harriet Cox Senior Design Competition Projects

Part of every UNLV engineering student’s academic experience, the senior design project stimulates engineering innovation and entrepreneurship. Each student in their senior year chooses, plans, designs, and prototypes a product in this required element of the curriculum. A capstone to the student’s educational career, the senior design project encourages the student to use everything learned in the engineering program to create a practical, real world solution to an engineering challenge. The senior design competition helps focus the senior students in increasing the quality and potential for commercial application for their design projects. Judges from local industry evaluate the projects on …

Tcp Server And Client: Bookstore Enquiry, Fawaz Bukhowa

Student Scholar Symposium Abstracts and Posters

An application called "Bookstore Enquiry", and it is implemented in Java using TCP client-server program. It contains two programs; one program is called "Server" and another one is called "Client". In this application, the 'server' maintains information about books and for each book it stores information like 'BookId', 'BookName', 'BookEdition', 'AvailableStock', 'UnitPrice', 'Discount'. This application works in such a way that, the server runs indefinitely and waits for client requests. The Client will accept the BookId & BookName from console and send it to server. If the server finds any books that matches with sent details, then it shows "BOOK …

Using Deep Learning To Forecast Spatiotemporal Crime Patterns, Shuzhan Fan

LSU Doctoral Dissertations

The distributional patterns of crime occurrences are closely related to their spatial, temporal, and environmental contexts. It has been a hot topic for researchers and crime analysts to discover such complex relationships in order to forecast crime, both spatially and temporally. Many factors play a role in the occurrences of crimes. Conventional crime forecasting research has primarily relied on historical crime records and socioeconomic data, while ignoring the rich social media and other environmental context data. The large volume of data requires a more appropriate forecasting framework with the ability to take in massive multimodal data and possibly achieve better …

Gmaim: An Analytical Pipeline For Microrna Splicing Profiling Using Generative Model, Kan Liu

School of Computing: Dissertations, Theses, and Student Research

MicroRNAs (miRNAs) are a class of short (~22 nt) single strand RNA molecules predominantly found in eukaryotes. Being involved in many major biological processes, miRNAs can regulate gene expression by targeting mRNAs to facilitate their degradation or translational inhibition. The imprecise splicing of miRNA splicing which introduces severe variability in terms of sequences of miRNA products and their corresponding downstream gene expression regulation. For example, to study biogenesis of miRNAs, usually, biologists can deplete a gene in the miRNA biogenesis pathway and study the change of miRNA sequences, which can cause impression of miRNAs. Although high-throughput sequencing technologies such as …

Historical Effects Of Electronic Interfaces, G James Mitchell

Publications and Research

Electronic interfaces are a primary tool for most professional and personal communication currently happening. Electronics, like the human mind, are limited by the understanding of executing will, or commands. This can be characterized as “interface limitations” of digital technology. Identifying this bottleneck in technological development has been critical in historical changes to both hardware and software technology. Recent medical research examines a novel user interface to reduce task load. I hypothesize, interface developments that take cues from nonverbal human communication enhance and sustain the significance of those technologies in society. By examining pivotal moments of historical technology we can identify …

Bots, Bias And Big Data: Artificial Intelligence, Algorithmic Bias And Disparate Impact Liability In Hiring Practices, Mckenzie Raub

Arkansas Law Review

This discussion is divided into three basic parts. Part one will provide a brief overview of artificial intelligence technology, its societal implications, and use emerging uses in hiring. Part two will discuss the potential for Title VII disparate impact arising from the use of artificial intelligence in hiring. Finally, part three will discuss proposed solutions to the challenges associated with the use of artificial intelligence technology, ultimately advocating for an approach that involves careful selection of the artificial intelligence program and balancing the use of artificial intelligence technology with human intuition.

Facepet: Enhancing Bystanders' Facial Privacy With Smart Wearables/Internet Of Things, Alfredo J. Perez, Sherali Zeadally, Luis Y. Matos Garcia, Jaouad A. Mouloud, Scott Griffith

Information Science Faculty Publications

Given the availability of cameras in mobile phones, drones and Internet-connected devices, facial privacy has become an area of major interest in the last few years, especially when photos are captured and can be used to identify bystanders’ faces who may have not given consent for these photos to be taken and be identified. Some solutions to protect facial privacy in photos currently exist. However, many of these solutions do not give a choice to bystanders because they rely on algorithms that de-identify photos or protocols to deactivate devices and systems not controlled by bystanders, thereby being dependent on the …

Enabling Auditing And Intrusion Detection Of Proprietary Controller Area Networks, Brent C. Stone

Theses and Dissertations

The goal of this dissertation is to provide automated methods for security researchers to overcome ‘security through obscurity’ used by manufacturers of proprietary Industrial Control Systems (ICS). `White hat' security analysts waste significant time reverse engineering these systems' opaque network configurations instead of performing meaningful security auditing tasks. Automating the process of documenting proprietary protocol configurations is intended to improve independent security auditing of ICS networks. The major contributions of this dissertation are a novel approach for unsupervised lexical analysis of binary network data flows and analysis of the time series data extracted as a result. We demonstrate the utility …

Facepet: Enhancing Bystanders’ Facial Privacy With Smart Wearables/Internet Of Things, Alfredo J. Perez, Sherali Zeadally, Luis Y. Matos Garcia, Jason A. Mouloud, Scott Griffith

Computer Science Faculty Publications

Given the availability of cameras in mobile phones, drones and Internet-connected devices, facial privacy has become an area of major interest in the last few years, especially when photos are captured and can be used to identify bystanders’ faces who may have not given consent for these photos to be taken and be identified. Some solutions to protect facial privacy in photos currently exist. However, many of these solutions do not give a choice to bystanders because they rely on algorithms that de-identify photos or protocols to deactivate devices and systems not controlled by bystanders, thereby being dependent on the …

Transient Solution Of An M/M/1 Retrial Queue With Reneging From Orbit, A. Azhagappan, E. Veeramani, W. Monica, K. Sonabharathi

Applications and Applied Mathematics: An International Journal (AAM)

In this paper, the transient behavior of an M/M/1 retrial queueing model is analyzed where the customers in the orbit possess the reneging behavior. There is no waiting room in the system for the arrivals. If the server is not free when the occurrence of an arrival, the arriving customer moves to the waiting group, known as orbit and retries for his service. If the server is idle when an arrival occurs (either coming from outside the queueing system or from the waiting group), the arrival immediately gets the service and leaves the system. Each individual customer in the orbit, …

An M^X/G(A,B)/1 Queue With Breakdown And Delay Time To Two Phase Repair Under Multiple Vacation, G. Ayyappan, M. Nirmala

Applications and Applied Mathematics: An International Journal (AAM)

In this paper, we consider an M^x /G(a,b)/1 queue with active breakdown and delay time to two phase repair under multiple vacation policy. A batch of customers arrive according to a compound Poisson process. The server serves the customers according to the “General Bulk Service Rule” (GBSR) and the service time follows a general (arbitrary) distribution. The server is unreliable and it may breakdown at any instance. As the result of breakdown, the service is suspended, the server waits for the repair to start and this waiting time is called as „delay time‟ and is assumed to follow general …

Analysis Of Batch Arrival Bulk Service Queue With Multiple Vacation Closedown Essential And Optional Repair, G. Ayyappan, T. Deepa

Applications and Applied Mathematics: An International Journal (AAM)

The objective of this paper is to analyze an queueing model with multiple vacation, closedown, essential and optional repair. Whenever the queue size is less than , the server starts closedown and then goes to multiple vacation. This process continues until at least customer is waiting in the queue. Breakdown may occur with probability when the server is busy. After finishing a batch of service, if the server gets breakdown with a probability , the server will be sent for repair. After the completion of the first essential repair, the server is sent to the second optional repair with probability …

Batch Arrival Bulk Service Queue With Unreliable Server, Second Optional Service, Two Different Vacations And Restricted Admissibility Policy, G. Ayyappan, R. Supraja

Applications and Applied Mathematics: An International Journal (AAM)

This paper is concerned with batch arrival queue with an additional second optional service to a batch of customers with dissimilar service rate where the idea of restricted admissibility of arriving batch of customers is also introduced. The server may take two different vacations (i) Emergency vacation-during service the server may go for vacation to an emergency call and after completion of the vacation, the server continues the remaining service to a batch of customers. (ii) Bernoulli vacation-after completion of first essential or second optional service, the server may take a vacation or may remain in the system to serve …

Learning-Based Analysis On The Exploitability Of Security Vulnerabilities, Adam Bliss

Computer Science and Computer Engineering Undergraduate Honors Theses

The purpose of this thesis is to develop a tool that uses machine learning techniques to make predictions about whether or not a given vulnerability will be exploited. Such a tool could help organizations such as electric utilities to prioritize their security patching operations. Three different models, based on a deep neural network, a random forest, and a support vector machine respectively, are designed and implemented. Training data for these models is compiled from a variety of sources, including the National Vulnerability Database published by NIST and the Exploit Database published by Offensive Security. Extensive experiments are conducted, including testing …

Computational Modeling Of Radiation Interactions With Molecular Nitrogen, Tyler Reese

Dissertations

The ability to detect radiation through identifying secondary effects it has on its surrounding medium would extend the range at which detections could be made and would be a valuable asset to many industries. The development of such a detection instrument requires an accurate prediction of these secondary effects. This research aims to improve on existing modeling techniques and help provide a method for predicting results for an affected medium in the presence of radioactive materials. A review of radioactivity and the interactions mechanisms for emitted particles as well as a brief history of the Monte Carlo Method and its …

The Rise Of Real-Time Retail Payments, Zhiling Guo

MITB Thought Leadership Series

TRANSACTING for just about anything using our mobile phones has become commonplace, and so many consumers will be intrigued to discover that after making a purchase it can still take longer for payment to reach a vendor’s bank account than it does for the purchased goods to be delivered.

Leveraging Artificial Intelligence To Capture The Singapore Rideshare Market, Pradeep Varakantham

MITB Thought Leadership Series

BIKE-SHARING programmes face many of the issues encountered by their counterparts in the carsharing world. But in Singapore, there are a number of factors that have a unique impact on the industry. These include the regulatory structure and the significant fines for those companies who do not abide by these regulations. When this is combined with the competitive nature of the industry in one of the world's most dynamic cities, it becomes clear that first movers who leverage machine learning and prediction will come to dominate the industry

The Analysis Of M/M/1 Queue With Working Vacation In Fuzzy Environment, G. Kannadasan, N. Sathiyamoorth

Applications and Applied Mathematics: An International Journal (AAM)

This study investigates the FM/FM/1 queue with working vacation. For this fuzzy queuing model, the researcher obtains some performance measure of interest such as the regular busy period, working vacation period, stationary queue length and waiting time. Finally, numerical results are presented to show the effects of system parameters.

Project Renew Worcester, Danni Yue, Amy Zhang, Jing Han, Omid Ashrafi, Yiming Xu

School of Professional Studies

n The client for this capstone project is RENEW Worcester which is a fledgling solar power project whose main goals are to bring renewable energy in the form of solar power into local, primarily low-income communities and are committed to the mission of making the transition off of fossil fuels to clean, renewable power. Based in Worcester, Massachusetts, they are a local chapter of Co-op Power which is a consumer-owned sustainable energy cooperative (co-op) made up of numerous different local chapters all over the New England area as well as the state of New York. The problem that we will …

Nba 2k, Joseph Saludo

ART 108: Introduction to Games Studies

The NBA 2K games have come a long way from an emerging basketball game to now becoming the biggest basketball game ever created. From its graphics, gameplay, community, and many more reasons why the game became so successful today, NBA 2K has evolved into the best basketball game by improving its overall structure every year-round.

The R Journal (December 2018) 10(2): Complete Issue, The R Foundation

The R Journal

Editorial, John Verzani

Contributed Research Articles

stplanr: A Package for Transport Planning, Robin Lovelace and Richard Ellison

The utiml Package: Multi-label Classification in R, Adriano Rivolli and Andre C. P. L. F. de Carvalho

rcss: R Package for Optimal Convex Stochastic Switching, Juri Hinz and Jeremy Yee

nsROC: An R package for Non-Standard ROC Curve Analysis, Sonia Pérez-Fernández, Pablo Martínez-Camblor, Peter Filzmoser, and Norberto Corral

addhaz: Contribution of Chronic Diseases to the Disability Burden Using R, Renata Tiene de Carvalho Yokota, Caspar WN Looman, Wilma Johanna Nusselder, Herman Van Oyen, and Geert Molenberghs

Snowboot: Bootstrap Methods for Network Inference, Yuzhou …

The Effect Of Incorporating End-User Customization Into Additive Manufacturing Designs, Jonathan D. Ashley

Graduate Theses and Dissertations

In the realm of additive manufacturing there is an increasing trend among makers to create designs that allow for end-users to alter them prior to printing an artifact. Online design repositories have tools that facilitate the creation of such artifacts. There are currently no rules for how to create a good customizable design or a way to measure the degree of customization within a design. This work defines three types of customizations found in additive manufacturing and presents three metrics to measure the degree of customization within designs based on the three types of customization. The goal of this work …

Deepsign: A Deep-Learning Architecture For Sign Language, Jai Amrish Shah

Computer Science and Engineering Theses - Archive

Sign languages are used by deaf people for communication. In sign languages, humans use hand gestures, body, facial expressions and movements to convey meaning. Humans can easily learn and understand sign languages, but automatic sign language recognition for machines is a challenging task. Using recent advances in the field of deep-learning, we introduce a fully automated deep-learning architecture for isolated sign language recognition. Our architecture tries to address three problems: 1) Satisfactory accuracy with limited data samples 2) Reducing chances of over-fitting when the data is limited 3) Automating recognition of isolated signs. Our architecture uses deep convolutional encoder-decoder architecture …

Classification Of Clinical Narratives Using Convolutional Neural Network, Nikit Rajiv Lonari

Computer Science and Engineering Theses - Archive

Patient safety is a key aspect for good consumer care. When an individual is hospitalized or receives medication the family wants the patient safety to be above all factors. For instance, a drug can do both either cure the disease or perhaps, give rise to an adverse event. A drug administered for an indicated condition has substantial power to reduce or cure a disease and further to prevent it from happening again in the future but at the risk of side effects. At present, there are several methods in patient safety and in particular in the area of signal detection …

Towards End-To-End Semi-Supervised Deep Learning For Drug Discovery, Xiaoyu Zhang

Computer Science and Engineering Theses - Archive

Observing the recent progress in Deep Learning, the employment of AI is surging to accelerate drug discovery and cut R&D costs in the last few years. However, the success of deep learning is attributed to large-scale clean high-quality labeled data, which is generally unavailable in drug discovery practices. In this thesis, we address this issue by proposing an end-to-end deep learning framework in a semi supervised learning fashion. That is said, the proposed deep learning approach can utilize both labeled and unlabeled data. While labeled data is of very limited availability, the amount of available unlabeled data is generally huge. …

Dwrelu : Double Weighted Rectifier Linear Unit An Activation Function With Trainable Scaling Parameter, Bhaskar Chandra Trivedi

Computer Science and Engineering Theses - Archive

Deep Neural Network have become very popular for computer vision application in recent years. At the same time, it remains important to understand the different implementation choices that need to be made when designing a neural network and to thoroughly investigate existing and novel alternatives for those choices. One of those choices is the activation function. The ReLU activation function is a widely used activation function. It discards all the values below zero and keeps the ones greater than zero. Variations such as Leaky ReLU and Parametric ReLU do not discard values, so that gradiants are nonzero for the entire …