Digital Commons Network^™

Parameterizing And Aggregating Activation Functions In Deep Neural Networks, Luke Benjamin Godfrey

Graduate Theses and Dissertations

The nonlinear activation functions applied by each neuron in a neural network are essential for making neural networks powerful representational models. If these are omitted, even deep neural networks reduce to simple linear regression due to the fact that a linear combination of linear combinations is still a linear combination. In much of the existing literature on neural networks, just one or two activation functions are selected for the entire network, even though the use of heterogenous activation functions has been shown to produce superior results in some cases. Even less often employed are activation functions that can adapt their …

Longitudinal Tracking Of Physiological State With Electromyographic Signals., Robert Warren Stallard

Electronic Theses and Dissertations

Electrophysiological measurements have been used in recent history to classify instantaneous physiological configurations, e.g., hand gestures. This work investigates the feasibility of working with changes in physiological configurations over time (i.e., longitudinally) using a variety of algorithms from the machine learning domain. We demonstrate a high degree of classification accuracy for a binary classification problem derived from electromyography measurements before and after a 35-day bedrest. The problem difficulty is increased with a more dynamic experiment testing for changes in astronaut sensorimotor performance by taking electromyography and force plate measurements before, during, and after a jump from a small platform. A …

A Framework For Cardio-Pulmonary Resuscitation (Cpr) Scene Retrieval From Medical Simulation Videos Based On Object And Activity Detection., Anju Panicker Madhusoodhanan Sathik

Electronic Theses and Dissertations

In this thesis, we propose a framework to detect and retrieve CPR activity scenes from medical simulation videos. Medical simulation is a modern training method for medical students, where an emergency patient condition is simulated on human-like mannequins and the students act upon. These simulation sessions are recorded by the physician, for later debriefing. With the increasing number of simulation videos, automatic detection and retrieval of specific scenes became necessary. The proposed framework for CPR scene retrieval, would eliminate the conventional approach of using shot detection and frame segmentation techniques. Firstly, our work explores the application of Histogram of Oriented …

Contractive Autoencoding For Hierarchical Temporal Memory And Sparse Distributed Representation Binding, Luke G. Boudreau

Theses

Hierarchical Temporal Memory is a brain inspired memory prediction framework modeled after the uniform structure and connectivity of pyramidal neurons found in the human neocortex. Similar to the neocortex, Hierarchical Temporal Memory processes spatiotemporal information for anomaly detection and prediction. A critical component in the Hierarchical Temporal Memory algorithm is the Spatial Pooler, which is responsible for processing feedforward data into sparse distributed representations.

This study addresses three fundamental research questions for Hierarchical Temporal Memory algorithms. What are the metrics for understanding the semantic content of sparse distributed representations? The semantic content and relationships between representations was visualized with uniqueness …

Assessing Classical And Expressive Aesthetics Of Web Pages Using Machine Learning, Ang Chen, Fiona Fui-Hoon Nah, Langtao Chen

Research Collection School Of Computing and Information Systems

Aesthetics plays a key role in web design. However, most websites are developed based on designers’ "inspirations" or "educated guesses" (Liu, 2003). While perceptions of aesthetics are intuitive abilities of humankind, the underlying principles for assessing aesthetics are not well understood. In this research, we propose using machine learning techniques to explore and more fully understand the patterns and underlying principles of aesthetics. We propose using machine learning techniques to develop predictive models for two aesthetic dimensions – classical aesthetics and expressive aesthetics – as well as for overall aesthetics of web pages in order to evaluate the aesthetic quality …

Music Genre Classification With Neural Networks: An Examination Of Several Impactful Variables, Jingqing Yang

Computer Science Honors Theses

There have been several attempts to classify music with content-based machine learning approaches. Most of these projects followed a similar procedure with a Deep Belief Network. In this project, we examined the performance of convolutional neural networks (CNN) and recurrent neural networks (RNN) as well as other components of a classification architecture, such as the choice of dataset, pre-processing techniques, and the sample size. Under a controlled environment, we discovered that the most successful architecture was a Mel-spectrogram combined with a CNN. Although our results fell behind the state-of-the-art performance, we outperform other music classification studies that use a CNN …

Evaluating Sequence Discovery Systems In An Abstraction-Aware Manner, Eoin Rogers, Robert J. Ross, John D. Kelleher

Conference papers

Activity discovery is a challenging machine learning problem where we seek to uncover new or altered behavioural patterns in sensor data. In this paper we motivate and introduce a novel approach to evaluating activity discovery systems. Pre-annotated ground truths, often used to evaluate the performance of such systems on existing datasets, may exist at different levels of abstraction to the output of the output produced by the system. We propose a method for detecting and dealing with this situation, allowing for useful ground truth comparisons. This work has applications for activity discovery, and also for related fields. For example, it …

Learning From Mutants: Using Code Mutation To Learn And Monitor Invariants Of A Cyber-Physical System, Yuqi Chen, Christopher M. Poskitt, Jun Sun

Research Collection School Of Computing and Information Systems

Cyber-physical systems (CPS) consist of sensors, actuators, and controllers all communicating over a network; if any subset becomes compromised, an attacker could cause significant damage. With access to data logs and a model of the CPS, the physical effects of an attack could potentially be detected before any damage is done. Manually building a model that is accurate enough in practice, however, is extremely difficult. In this paper, we propose a novel approach for constructing models of CPS automatically, by applying supervised machine learning to data traces obtained after systematically seeding their software components with faults ("mutants"). We demonstrate the …

A Qualitative Research On Marketing And Sales In The Artificial Intelligence Age, Yin Yang, Keng Siau

Research Collection School Of Computing and Information Systems

The age of artificial intelligence is here! Artificial Intelligence, robotics, machine learning, and automation are impacting the field of marketing and sales in an unprecedented way. In this study, the qualitative research methodology will be used to better understand the revolution and evolution of marketing and sales field in the AI age. Multiple case studies will be performed in various marketing and sales units in different organizations. This research is of value to both academics and practitioners as it aims to provide a detailed analysis and documentation of the changes in marketing and sales functionalities and job markets as AI …

Patterns, Collaboration, Practice: Algorithms As Editing For Historic Periodicals, Elizabeth Lorang

University of Nebraska-Lincoln Libraries: Presentations

This presentation positions my recent work on the algorithmic “discovery” of poetic material in historic newspapers within the contexts of my various roles as an editor of periodical literature and also consider how duplicative processes and algorithms encode principles and values and function as editorial acts. Ultimately, I hope to pose a range of questions to prompt discussion around the place (or not) of machine learning in identifying and selecting texts and bodies of work; what ideas we’re actually exploring/are able to explore when we enlist technology in stages of this work; and the stakes of these activities, whether human …

Detecting Suicide Risk From Wristworn Activity Tracker Data Using Machine Learning Approaches, Pallavi Atluri

Electrical Engineering Theses

Suicide is a prevalent cause of death worldwide and depression is a primary concern of many suicidal acts. It is possible that an individual during depression never has any suicidal thoughts at all. On the other hand, some individuals in stable condition with no apparent symptoms of depression feel urges to commit suicide (suicidal ideation). Many such individuals never let anyone know what they are feeling or planning. Suicidal ideation considered an important precursor to suicidal acts.

Detecting the suicide risk in individuals with mood disorders is a major challenge. The current clinical practice to assess suicide risk in these …

Comparative Study Of Deep Learning Models For Network Intrusion Detection, Brian Lee, Sandhya Amaresh, Clifford Green, Daniel Engels

SMU Data Science Review

In this paper, we present a comparative evaluation of deep learning approaches to network intrusion detection. A Network Intrusion Detection System (NIDS) is a critical component of every Internet connected system due to likely attacks from both external and internal sources. A NIDS is used to detect network born attacks such as Denial of Service (DoS) attacks, malware replication, and intruders that are operating within the system. Multiple deep learning approaches have been proposed for intrusion detection systems. We evaluate three models, a vanilla deep neural net (DNN), self-taught learning (STL) approach, and Recurrent Neural Network (RNN) based Long Short …

Cognitive Virtual Admissions Counselor, Kumar Raja Guvindan Raju, Cory Adams, Raghuram Srinivas

SMU Data Science Review

Abstract. In this paper, we present a cognitive virtual admissions counselor for the Master of Science in Data Science program at Southern Methodist University. The virtual admissions counselor is a system capable of providing potential students accurate information at the time that they want to know it. After the evaluation of multiple technologies, Amazon’s LEX was selected to serve as the core technology for the virtual counselor chatbot. Student surveys were leveraged to collect and generate training data to deploy the natural language capability. The cognitive virtual admissions counselor platform is currently capable of providing an end-to-end conversational dialog to …

Using Data Visualization To Inform Machine Learning Approaches, Eric N. Zielonka, Aleksandr Fritz

STEM Student Research Symposium Posters

Machine learning with big data is a complicated task to tackle. Using data visualizations, one can find trends, anomalies, and patterns to help select the appropriate approach to the problem in machine learning. Using 2D visualizations, we’ve displayed flight data on interactive maps, visualizing density and property changes in an area. We’ve also used frequency histograms to view the quantitative properties of each point to look for trends. Using scatterplots, anomalies in data collection were found. Other plots confirmed previously found trends and initial thoughts about the data. These visualizations helped inform a machine learning approach to our problem and …

A Convolutional Neural Network Model For Species Classification Of Camera Trap Images, Annie Casey

Mathematics Undergraduate Theses

The overall purpose of this study was to automate the manual process of tagging species found in camera trap images using machine learning. The basic design of this study was to implement a Convolutional Neural Network model in Python using the Keras and Tensorflow modules that learn to recognize patterns in images in order to classify what species is in a given image and to label it accordingly. Results of the analysis highlight the importance of a large sample size, the degree of accuracy according to various arguments in the model, effectiveness of multiple layers that include Max Pooling, and …

Exploring The Use Of Hierarchal Statistical Analysis And Deep Neural Networks To Detect And Mitigate Covert Timing Channels, Omar Darwish

Dissertations

Covert timing channels provide a mechanism to transmit unauthorized information across different processes. It utilizes the inter-arrival times between the transmitted packets to hide the communicated data. It can be exploited in a variety of malevolent scenarios such as leaking military secrets, trade secrets, and other forms of Intellectual Property (IP). They can be also used as a vehicle to attack existing computing systems to disseminate software viruses or worms while bypassing firewalls, intrusion detection and protection systems, and application filters. Therefore, the detection and mitigation of covert channels is a key issue in modern Information Technology (IT) infrastructure. Many …

A Holistic Computational Approach To Boosting The Performance Of Protein Search Engines, Majdi Ahmad Mosa Maabreh

Dissertations

Despite availability of several proteins search engines, due to the increasing amounts of MS/MS data and database sizes, more efficient data analysis and reduction methods are important. Improving accuracy and performance of protein identification is a main goal in the community of proteomic research. In this research, a holistic solution for improvement in search performance is developed.

Most current search engines apply the SEQUEST style of searching protein databases to define MS/MS spectra. SEQUEST involves three main phases: (i) Indexing the protein databases, (ii) Matching and Ranking the MS/MS spectra and (iii) Filtering the matches and reporting the final proteins. …

Using Computational Text Classification For Qualitative Research And Evaluation In Extension, Justin G. Smith, Reid Tissing

Journal of Extension

This article introduces a process for computational text classification that can be used in a variety of qualitative research and evaluation settings. The process leverages supervised machine learning based on an implementation of a multinomial Bayesian classifier. Applied to a community of inquiry framework, the algorithm was used to identify evidence of cognitive presence, social presence, and teaching presence in the text contributions (44,000 unique posts) of more than 4,000 participants in an online environmental education course. Results indicate that computational text classification can significantly reduce labor costs and can help Extension research faculty scale, accelerate, and ensure reproducibility of …

A Comparison Of Machine Learning Techniques For Taxonomic Classification Of Teeth From The Family Bovidae, Gregory J. Matthews, Juliet K. Brophy, Maxwell Luetkemeier, Hongie Gu, George K. Thiruvathukal

Mathematics and Statistics: Faculty Publications and Other Works

This study explores the performance of machine learning algorithms on the classification of fossil teeth in the Family Bovidae. Isolated bovid teeth are typically the most common fossils found in southern Africa and they often constitute the basis for paleoenvironmental reconstructions. Taxonomic identification of fossil bovid teeth, however, is often imprecise and subjective. Using modern teeth with known taxons, machine learning algorithms can be trained to classify fossils. Previous work by Brophy et al. [Quantitative morphological analysis of bovid teeth and implications for paleoenvironmental reconstruction of plovers lake, Gauteng Province, South Africa, J. Archaeol. Sci. 41 (2014), pp. …

Opportunity Identification For New Product Planning: Ontological Semantic Patent Classification, Farshad Madani

Dissertations and Theses

Intelligence tools have been developed and applied widely in many different areas in engineering, business and management. Many commercialized tools for business intelligence are available in the market. However, no practically useful tools for technology intelligence are available at this time, and very little academic research in technology intelligence methods has been conducted to date.

Patent databases are the most important data source for technology intelligence tools, but patents inherently contain unstructured data. Consequently, extracting text data from patent databases, converting that data to meaningful information and generating useful knowledge from this information become complex tasks. These tasks are currently …

Hierarchical Random Boolean Network Reservoirs, Sai Kiran Cherupally

Dissertations and Theses

Reservoir Computing (RC) is an emerging Machine Learning (ML) paradigm. RC systems contain randomly assembled computing devices and can be trained to solve complex temporal tasks. These systems are computationally cheaper to train than other ML paradigms such as recurrent neural networks, and they can also be trained to solve multiple tasks simultaneously. Further, hierarchical RC systems with fixed topologies, were shown to outperform monolithic RC systems by up to 40% when solving temporal tasks. Although the performance of monolithic RC networks was shown to improve with increasing network size, building large monolithic networks may be challenging, for example because …

Secrecy In Educational Practices: Enacting Nested Black Boxes In Cheating And Deception Detection Systems, Jo Ann Oravec

Secrecy and Society

This paper covers secrecy from the vantage point of recent technological initiatives designed to detect cheating and deception in educational contexts as well as to monitor off-campus social media speech code violations. Many of these systems are developed and implemented by third-party corporate entities who claim practices to be proprietary and secret. The outsourcers involved in these efforts have provided one level of secrecy and educational administrators involved yet another level, thus constructing "nested black boxes." Also discussed in this paper is the “paranoid style” of administration, often supported by the surveillance and construction of rosters of potential non-conformists, such …

Machine Learning From Fetal Flow Waveforms To Predict Adverse Perinatal Outcomes: A Study Protocol, Zahra Hoodbhoy, Babar Hasan, Fyezah Jehan, Bart Bijnens, Devyani Chowdhury

Department of Paediatrics and Child Health

Background: In Pakistan, stillbirth rates and early neonatal mortality rates are amongst the highest in the world. The aim of this study is to provide proof of concept for using a computational model of fetal haemodynamics, combined with machine learning. This model will be based on Doppler patterns of the fetal cardiovascular, cerebral and placental flows with the goal to identify those fetuses at increased risk of adverse perinatal outcomes such as stillbirth, perinatal mortality and other neonatal morbidities.
Methods: This will be prospective one group cohort study which will be conducted in Ibrahim Hyderi, a peri-urban settlement in south …

Gradient Estimation For Attractor Networks, Thomas Flynn

Dissertations, Theses, and Capstone Projects

It has been hypothesized that neural network models with cyclic connectivity may be more powerful than their feed-forward counterparts. This thesis investigates this hypothesis in several ways. We study the gradient estimation and optimization procedures for several variants of these networks. We show how the convergence of the gradient estimation procedures are related to the properties of the networks. Then we consider how to tune the relative rates of gradient estimation and parameter adaptation to ensure successful optimization in these models. We also derive new gradient estimators for stochastic models. First, we port the forward sensitivity analysis method to the …

Object Localization, Segmentation, And Classification In 3d Images, Allan Zelener

Dissertations, Theses, and Capstone Projects

We address the problem of identifying objects of interest in 3D images as a set of related tasks involving localization of objects within a scene, segmentation of observed object instances from other scene elements, classifying detected objects into semantic categories, and estimating the 3D pose of detected objects within the scene. The increasing availability of 3D sensors motivates us to leverage large amounts of 3D data to train machine learning models to address these tasks in 3D images. Leveraging recent advances in deep learning has allowed us to develop models capable of addressing these tasks and optimizing these tasks jointly …

A Machine Learning Algorithm For Identifying And Tracking Bacteria In Three Dimensions Using Digital Holographic Microscopy, Manuel Bedrossian, Marwan El-Kholy, Daniel Neamati, Jay Nadeau

Physics Faculty Publications and Presentations

Digital Holographic Microscopy (DHM) is an emerging technique for three-dimensional imaging of microorganisms due to its high throughput and large depth of field relative to traditional microscopy techniques. While it has shown substantial success for use with eukaryotes, it has proven challenging for bacterial imaging because of low contrast and sources of noise intrinsic to the method (e.g. laser speckle). This paper describes a custom written MATLAB routine using machine-learning algorithms to obtain three-dimensional trajectories of live, lab-grown bacteria as they move within an essentially unrestrained environment with more than 90% precision. A fully annotated version of the software used …

Increasing Our Vision For 21st-Century Digital Libraries, Elizabeth M. Lorang, Leen-Kiat Soh

University of Nebraska-Lincoln Libraries: Presentations

This presentation

1. Reads digital library interfaces—or their "main door" interfaces—as glimpses into what we have thus far valued in the development of digital libraries
2. Frames a visual way of thinking about textual materials
3. Introduces the work of our research team—where we are now, and where we're headed
4. Draws some connections between the parts

This presentation is very much a look into thinking in process and work in progress and proposes the following ideas:

1. As a community, we can do much more with the digital images we're creating of textual materials than we've heretofore done.
2. We aspire to have additional layers …

Deep Learning Of 2-D Images Representing N-D Data In General Line Coordinates, Dmytro Dovhalets, Boris Kovalerchuk, Szilárd Vajda, Răzvan Andonie

Computer Science Faculty Scholarship

While knowledge discovery and n-D data visualization procedures are often efficient, the loss of information, occlusion, and clutter continue to be a challenge. General Line Coordinates (GLC) is a rather new technique to deal with such artifacts. GLC-Linear, which is one of the methods in GLC, allows transforming n-D numerical data to their visual representation as polylines losslessly. The method proposed in this paper uses these 2-D visual representations as input to Convolutional Neural Network (CNN) classifiers. The obtained classification accuracies are close to the ones obtained by other machine learning algorithms. The main benefit of the method is the …

Automating The Crowd-Mapping Workflow With Deep Learning, Lasith Niroshan

Theses

Maintaining updated maps in an ever-changing built environment is important for supporting modern society in many ways. The usage of online crowdsourced maps in particular has gained importance in a wide range of recent location-based applications (route planning/navigation, urban planning, real estate, tourism, etc). However, both traditional map production methods and updating today’s online maps suffer from early obsolescence due to their largely manual map production/update workflows. Significant research efforts have focused on refining techniques to identify changes in raster satellite images, aiming to improve and streamline map production processes. Concurrently, the surge in Internet usage has led to a …

Unsupervised Machine Learning Account Of Magnetic Transitions In The Hubbard Model, Kelvin Ch'ng, Nick Vazquez, Ehsan Khatami

Faculty Publications

We employ several unsupervised machine learning techniques, including autoencoders, random trees embedding, and t-distributed stochastic neighboring ensemble (t-SNE), to reduce the dimensionality of, and therefore classify, raw (auxiliary) spin configurations generated, through Monte Carlo simulations of small clusters, for the Ising and Fermi-Hubbard models at finite temperatures. Results from a convolutional autoencoder for the three-dimensional Ising model can be shown to produce the magnetization and the susceptibility as a function of temperature with a high degree of accuracy. Quantum fluctuations distort this picture and prevent us from making such connections between the output of the autoencoder and …