Improve Image Classification Using Data Augmentation And Neural Networks, 2019 Southern Methodist University
Improve Image Classification Using Data Augmentation And Neural Networks, Shanqing Gu, Manisha Pednekar, Robert Slater
SMU Data Science Review
In this paper, we present how to improve image classification by using data augmentation and convolutional neural networks. Model overfitting and poor performance are common problems in applying neural network techniques. Approaches to bring intra-class differences down and retain sensitivity to the inter-class variations are important to maximize model accuracy and minimize the loss function. With CIFAR-10 public image dataset, the effects of model overfitting were monitored within different model architectures in combination of data augmentation and hyper-parameter tuning. The model performance was evaluated with train and test accuracy and loss, characteristics derived from the confusion matrices, and visualizations of ...
Development Of A Statistical Shape-Function Model Of The Implanted Knee For Real-Time Prediction Of Joint Mechanics, 2019 Boise State University
Development Of A Statistical Shape-Function Model Of The Implanted Knee For Real-Time Prediction Of Joint Mechanics, Kalin Gibbons
Boise State University Theses and Dissertations
Outcomes of total knee arthroplasty (TKA) are dependent on surgical technique, patient variability, and implant design. Non-optimal design or alignment choices may result in undesirable contact mechanics and joint kinematics, including poor joint alignment, instability, and reduced range of motion. Implant design and surgical alignment are modifiable factors with potential to improve patient outcomes, and there is a need for robust implant designs that can accommodate patient variability. Our objective was to develop a statistical shape-function model (SFM) of a posterior stabilized implant knee to instantaneously predict output mechanics in an efficient manner. Finite element methods were combined with Latin ...
Adaptation Of A Deep Learning Algorithm For Traffic Sign Detection, 2019 The University of Western Ontario
Adaptation Of A Deep Learning Algorithm For Traffic Sign Detection, Jose Luis Masache Narvaez
Electronic Thesis and Dissertation Repository
Traffic signs detection is becoming increasingly important as various approaches for automation using computer vision are becoming widely used in the industry. Typical applications include autonomous driving systems, mapping and cataloging traffic signs by municipalities. Convolutional neural networks (CNNs) have shown state of the art performances in classification tasks, and as a result, object detection algorithms based on CNNs have become popular in computer vision tasks. Two-stage detection algorithms like region proposal methods (R-CNN and Faster R-CNN) have better performance in terms of localization and recognition accuracy. However, these methods require high computational power for training and inference that make ...
An Efficient Framework For The Stochastic Verification Of Computation And Communication Systems Using Emerging Technologies, Zhen Zhang
Funded Research Records
No abstract provided.
The Perils And Promises Of Artificial General Intelligence, 2019 Notre Dame Law School
The Perils And Promises Of Artificial General Intelligence, Brian S. Haney
Journal of Legislation
No abstract provided.
Machine Learning With Multi-Class Regression And Neural Networks: Analysis And Visualization Of Crime Data In Seattle, 2019 Seattle Pacific University
Machine Learning With Multi-Class Regression And Neural Networks: Analysis And Visualization Of Crime Data In Seattle, Erkin David George
This article examines the implications of machine learning algorithms and models, and the significance of their construction when investigating criminal data. It uses machine learning models and tools to store, clean and analyze data that is fed into a machine learning model. This model is then compared to another model to test for accuracy, biases and patterns that are detected in between the experiments. The data was collected from data.seattle.gov and was published by the City of Seattle Data Portal and was accessed on September 17, 2018. This research will be looking into how machine learning models can ...
Transient Flow Analysis Of A Closing Blowout Preventer Using Computational Fluid Dynamics (Cfd), 2019 Louisiana State University
Transient Flow Analysis Of A Closing Blowout Preventer Using Computational Fluid Dynamics (Cfd), Daniel Barreca
LSU Master's Theses
Reliability of blowout preventers (BOPs) is crucial for drilling and production operations. Erosion of BOP components and hydrodynamic forces on rams may cause failure of BOP elements to seal the well. Transient computational fluid dynamics (CFD) simulations of fluids within the wellbore and BOP offer quantitative and qualitative data related to this reliability during the closure of various BOP components. Since limited research has been published in transient CFD simulations of closing BOPs, this thesis discusses challenges and solutions to simulating closing blowout preventers. Single component fluids are simulated through several BOP geometries such as annular preventers, pipe rams, and ...
Fifth Aeon – A.I Competition And Balancer, 2019 California Polytechnic State University, San Luis Obispo
Fifth Aeon – A.I Competition And Balancer, William M. Ritson
Master's Theses and Project Reports
Collectible Card Games (CCG) are one of the most popular types of games in both digital and physical space. Despite their popularity, there is a great deal of room for exploration into the application of artificial intelligence in order to enhance CCG gameplay and development. This paper presents Fifth Aeon a novel and open source CCG built to run in browsers and two A.I applications built upon Fifth Aeon. The first application is an artificial intelligence competition run on the Fifth Aeon game. The second is an automatic balancing system capable of helping a designer create new cards that ...
Cloneless: Code Clone Detection Via Program Dependence Graphs With Relaxed Constraints, 2019 California Polytechnic State University, San Luis Obispo
Cloneless: Code Clone Detection Via Program Dependence Graphs With Relaxed Constraints, Thomas J. Simko
Master's Theses and Project Reports
Code clones are pieces of code that have the same functionality. While some clones may structurally match one another, others may look drastically different. The inclusion of code clones clutters a code base, leading to increased costs through maintenance. Duplicate code is introduced through a variety of means, such as copy-pasting, code generated by tools, or developers unintentionally writing similar pieces of code. While manual clone identification may be more accurate than automated detection, it is infeasible due to the extensive size of many code bases. Software code clone detection methods have differing degree of success based on the analysis ...
A Ulysses Pact With Artificial Systems. How To Deliberately Change The Objective Spirit With Cultured Ai, 2019 University of Siegen, Germany
A Ulysses Pact With Artificial Systems. How To Deliberately Change The Objective Spirit With Cultured Ai, Bruno Gransche
Computer Ethics - Philosophical Enquiry (CEPE) Proceedings
The article introduces a concept of cultured technology, i.e. intelligent systems capable of interacting with humans and showing (or simulating) manners, of following customs and of socio-sensitive considerations. Such technologies might, when deployed on a large scale, influence and change the realm of human customs, traditions, standards of acceptable behavior, etc. This realm is known as the "objective spirit" (Hegel), which usually is thought of as being historically changing but not subject to deliberate human design. The article investigates the question of whether the purposeful design of interactive technologies (as cultured technologies) could enable us to shape modes of ...
A Machine Learning Approach To Predicting Alcohol Consumption In Adolescents From Historical Text Messaging Data, Adrienne Bergh
Computational and Data Sciences (MS) Theses
Techniques based on artificial neural networks represent the current state-of-the-art in machine learning due to the availability of improved hardware and large data sets. Here we employ doc2vec, an unsupervised neural network, to capture the semantic content of text messages sent by adolescents during high school, and encode this semantic content as numeric vectors. These vectors effectively condense the text message data into highly leverageable inputs to a logistic regression classifier in a matter of hours, as compared to the tedious and often quite lengthy task of manually coding data. Using our machine learning approach, we are able to train ...
Analytical And Numerical Modeling Of Cavity Expansion In Anisotropic Poroelastoplastic Soil, 2019 Louisiana State University and Agricultural and Mechanical College
Analytical And Numerical Modeling Of Cavity Expansion In Anisotropic Poroelastoplastic Soil, Kai Liu
LSU Doctoral Dissertations
Cavity expansion/contraction problems have attracted intensive attentions over the past several decades due to its versatile applications, such as the interpretation of pressuremeter/piezocone penetration testing results and the modelling of pile installation/tunnel excavation in civil engineering, and the prediction of critical mud pressure required to maintain the wellbore stability in petroleum engineering. Despite the fact that various types of constitutive models have been covered in the literature on this subject, the soils and/or rocks were usually treated as isotropic geomaterials.
In recognition of the above fact, this research makes a substantial extension of the fundamental cavity ...
University Of Rhode Island Course Information Assistant, 2019 University of Rhode Island
University Of Rhode Island Course Information Assistant, Daniel Gauthier
Senior Honors Projects
Personal voice-interactive systems have become ubiquitous in daily life. There are many of these digital assistants such as Siri, Alexa, and Google Assistant. The chances are high you have access to one right now. This technology has reached a point where the context of a conversation can be maintained, which is a vast improvement over earlier technology. Interactions without conversational context can limit interactions greatly and this was the case for previous digital assistants. Every time someone would say something to an assistant, it was like they were constantly changing operators on a customer service line. The assistants can now ...
Clustering Heterogeneous Autism Spectrum Disorder Data., 2019 University of Louisville
Clustering Heterogeneous Autism Spectrum Disorder Data., Mariem Boujelbene
Electronic Theses and Dissertations
Autism spectrum disorder (ASD) is a developmental disorder that affects communication and behavior. Several studies have been conducted in the past years to develop a better understanding of the disease and therefore a better diagnosis and a better treatment by analyzing diverse data sets consisting of behavioral surveys and tests, phenotype description, and brain imagery. However, data analysis is challenged by the diversity, complexity and heterogeneity of patient cases and by the need for integrating diverse data sets to reach a better understanding of ASD. The aim of our study is to mine homogeneous groups of patients from a heterogeneous ...
Receptive Fields Optimization In Deep Learning For Enhanced Interpretability, Diversity, And Resource Efficiency., 2019 University of Louisville
Receptive Fields Optimization In Deep Learning For Enhanced Interpretability, Diversity, And Resource Efficiency., Babajide Odunitan Ayinde
Electronic Theses and Dissertations
In both supervised and unsupervised learning settings, deep neural networks (DNNs) are known to perform hierarchical and discriminative representation of data. They are capable of automatically extracting excellent hierarchy of features from raw data without the need for manual feature engineering. Over the past few years, the general trend has been that DNNs have grown deeper and larger, amounting to huge number of final parameters and highly nonlinear cascade of features, thus improving the flexibility and accuracy of resulting models. In order to account for the scale, diversity and the difficulty of data DNNs learn from, the architectural complexity and ...
A Machine Learning Approach To Network Intrusion Detection System Using K Nearest Neighbor And Random Forest, 2019 University of Southern Mississippi
A Machine Learning Approach To Network Intrusion Detection System Using K Nearest Neighbor And Random Forest, Ilemona S. Atawodi
The evolving area of cybersecurity presents a dynamic battlefield for cyber criminals and security experts. Intrusions have now become a major concern in the cyberspace. Different methods are employed in tackling these threats, but there has been a need now more than ever to updating the traditional methods from rudimentary approaches such as manually updated blacklists and whitelists. Another method involves manually creating rules, this is usually one of the most common methods to date.
A lot of similar research that involves incorporating machine learning and artificial intelligence into both host and network-based intrusion systems recently. Doing this originally presented ...
Using Transfer Learning In Network Markets, 2019 The Graduate Center, City University of New York
Using Transfer Learning In Network Markets, Kai Cai
All Dissertations, Theses, and Capstone Projects
Mechanism design is the sub-field of microeconomics and game theory, which considers agents have their own private information and are self-interested and tries to design systems that can produce desirable outcomes. In recent years, with the development of internet and electronic markets, mechanism design has become an important research field in computer science. This work has largely focused on single markets. In the real world, individual markets tend to connect to other markets and form a big “network market”, where each market occupies a node in the network and connections between markets reflect constraints on traders in the markets. So ...
Modeling And Counteracting Exposure Bias In Recommender Systems., 2019 University of Louisville
Modeling And Counteracting Exposure Bias In Recommender Systems., Sami Khenissi
Electronic Theses and Dissertations
Recommender systems are becoming widely used in everyday life. They use machine learning algorithms which learn to predict our preferences and thus influence our choices among a staggering array of options online, such as movies, books, products, and even news articles. Thus what we discover and see online, and consequently our opinions and decisions, are becoming increasingly affected by automated predictions made by learning machines. Similarly, the predictive accuracy of these learning machines heavily depends on the feedback data, such as ratings and clicks, that we provide them. This mutual influence can lead to closed-loop interactions that may cause unknown ...
Forecasting Model For Disease Propensity Using Ehr Data, 2019 The University of San Francisco
Forecasting Model For Disease Propensity Using Ehr Data, Soodabeh Sarafrazi, Omar Sharif, Matthew Domingo, Jie Han, Michael Chang, Omid Khazaie, Anil Kemisetti
Creative Activity and Research Day - CARD
Many diseases such as diabetes and cardiovascular diseases are actionable, i.e. they are preventable by early intervention. One to two years of early warning would represent a huge advance in dealing with these conditions and could help prevent further complications such as heart disease, stroke, kidney failure, blindness, and amputation. In this project, we are developing an extensible condition forecasting model to assess the risk of diabetes and heart problems in patients in advance. Using TensorFlow, Elastic MapReduce (EMR), and AWS Sagemaker, we are training a Wide and Deep Neural Network on a dataset of more than 170 million ...
Deep Learning, Medical Physics And Cargo Cult Science., 2019 University of San Francisco
Deep Learning, Medical Physics And Cargo Cult Science., Miguel Romero Phd, Gilmer Valdes Phd, Timothy Solberg Phd, Yannet Interian Phd
Creative Activity and Research Day - CARD
Deep learning algorithms have become widely popular, with considerable success in fields where datasets have hundreds of thousands or million points. As deep learning is increasingly applied to the fields of medical physics and radiation oncology, a reasonable question follows: are these techniques the best approach, given the unique conditions in our field? In this study, we investigate the dependence of dataset size on the performance of deep learning algorithms compared with more traditional radiomics-based methods.