Physical Sciences and Mathematics Commons^™

Evaluation Of Supervised Machine Learning For Classifying Video Traffic, Farrell R. Taylor

CCE Theses and Dissertations

Operational deployment of machine learning based classifiers in real-world networks has become an important area of research to support automated real-time quality of service decisions by Internet service providers (ISPs) and more generally, network administrators. As the Internet has evolved, multimedia applications, such as voice over Internet protocol (VoIP), gaming, and video streaming, have become commonplace. These traffic types are sensitive to network perturbations, e.g. jitter and delay. Automated quality of service (QoS) capabilities offer a degree of relief by prioritizing network traffic without human intervention; however, they rely on the integration of real-time traffic classification to identify applications. Accordingly, …

Radical Recognition In Off-Line Handwritten Chinese Characters Using Non-Negative Matrix Factorization, Xiangying Shuai

Senior Projects Spring 2016

In the past decade, handwritten Chinese character recognition has received renewed interest with the emergence of touch screen devices. Other popular applications include on-line Chinese character dictionary look-up and visual translation in mobile phone applications. Due to the complex structure of Chinese characters, this classification task is not exactly an easy one, as it involves knowledge from mathematics, computer science, and linguistics.

Given a large image database of handwritten character data, the goal of my senior project is to use Non-Negative Matrix Factorization (NMF), a recent method for finding a suitable representation (parts-based representation) of image data, to detect specific …

Algorithmic Music Composition And Accompaniment Using Neural Networks, Daniel Wilton Risdon

Senior Projects Spring 2016

The goal of this project was to use neural networks as a tool for live music performance. Specifically, the intention was to adapt a preexisting neural network code library to work in Max, a visual programming language commonly used to create instruments and effects for electronic music and audio processing. This was done using ConvNetJS, a JavaScript library created by Andrej Karpathy.

Several neural network models were trained using a range of different training data, including music from various genres. The resulting neural network-based instruments were used to play brief pieces of music, which they used as input to create …

Applications Of Computational Geometry And Computer Vision, Joseph Lemley

All Master's Theses

Recent advances in machine learning research promise to bring us closer to the original goals of artificial intelligence. Spurred by recent innovations in low-cost, specialized hardware and incremental refinements in machine learning algorithms, machine learning is revolutionizing entire industries. Perhaps the biggest beneficiary of this progress has been the field of computer vision. Within the domains of computational geometry and computer vision are two problems: Finding large, interesting holes in high dimensional data, and locating and automatically classifying facial features from images. State of the art methods for facial feature classification are compared and new methods for finding empty hyper-rectangles …

An Extended Study On Addressing Defender Teamwork While Accounting For Uncertainty In Attacker Defender Games Using Iterative Dec-Mdps, Eric Shieh, Albert Xin Jiang, Amulya Yadav, Pradeep Varakantham, Milind Tambe

Research Collection School Of Computing and Information Systems

Multi-agent teamwork and defender-attacker security games are two areas that are currently receiving significant attention within multi-agent systems research. Unfortunately, despite the need for effective teamwork among multiple defenders, little has been done to harness the teamwork research in security games. The problem that this paper seeks to solve is the coordination of decentralized defender agents in the presence of uncertainty while securing targets against an observing adversary. To address this problem, we offer the following novel contributions in this paper: (i) New model of security games with defender teams that coordinate under uncertainty; (ii) New algorithm based on column …

Improving The Vector Auto Regression Technique For Time-Series Link Prediction By Using Support Vector Machine, Proceso L. Fernandez Jr, Jan Miles Co

Department of Information Systems & Computer Science Faculty Publications

Predicting links between the nodes of a graph has become an important Data Mining task because of its direct applications to biology, social networking, communication surveillance, and other domains. Recent literature in time-series link prediction has shown that the Vector Auto Regression (VAR) technique is one of the most accurate for this problem. In this study, we apply Support Vector Machine (SVM) to improve the VAR technique that uses an unweighted adjacency matrix along with 5 matrices: Common Neighbor (CN), Adamic-Adar (AA), Jaccard’s Coefficient (JC), Preferential Attachment (PA), and Research Allocation Index (RA). A DBLP dataset covering the years from …

Implementation Of An Air Supply Unit Control Scheme For The Uc2av (Unmanned Circulation Control Aerial Vehicle), Cameron Rosen

Electronic Theses and Dissertations

The expanded prevalence of Unmanned Aerial Vehicles (UAVs) in recent years has created many opportunities to research novel applications for their use, enabled by the reduced cost, mission flexibility, and reduced risk that small-scale unmanned platforms provide in comparison to larger aircraft. Despite the versatility of unmanned aviation, limitations on payload size and weight, fuel and power capacity, and takeoff and landing infrastructure can restrict UAV applications, and have created a need for lift augmenting technologies that can reduce the impact of these limitations. Circulation Control (CC) is an active flow technique that has been proven as a method for …

An Approach To Automatic Detection Of Suspicious Individuals In A Crowd, Satabdi Mukherjee

Dissertations and Theses

This paper describes an approach to identify individuals with suspicious objects in a crowd. It is based on a well-known image retrieval problem as applied to mobile visual search. In many cases, the process of building a hierarchical tree uses k-means clustering followed by geometric verification. However, the number of clusters is not known in advance, and sometimes it is randomly generated. This may lead to a congested clustering which can cause problems in grouping large real-time data. To overcome this problem we have applied the Indian Buffet stochastic process approach in this paper to the clustering problem. We present …

Automated Conjecturing Approach For Benzenoids, David Muncy

Theses and Dissertations

Benzenoids are graphs representing the carbon structure of molecules, defined by a closed path in the hexagonal lattice. These compounds are of interest to chemists studying existing and potential carbon structures. The goal of this study is to conjecture and prove relations between graph theoretic properties among benzenoids. First, we generate conjectures on upper bounds for the domination number in benzenoids using invariant-defined functions. This work is an extension of the ideas to be presented in a forthcoming paper. Next, we generate conjectures using property-defined functions. As the title indicates, the conjectures we prove are not thought of on our …

Enabling Carrier Collaboration Via Order Sharing Double Auction: A Singapore Urban Logistics Perspective, Handoko Stephanus Daniel, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

A recent exploratory study on the collaborative urban logistics in Singapore suggests that cost reduction and privacy preservation are two main drivers that would motivate the participation of carriers in consolidating their last mile deliveries. With Singapore's mild restrictions on the vehicle types or the time windows for the last-mile delivery, we believe that with proper technology in place, an Urban Consolidation Center like the Tenjin Joint Distribution System in Fukuoka Japan may be implemented to achieve cost reduction with some degree of privacy preservation. Participating carriers keep their respective private orders and have the option to get their remaining …

Techno-Apocalypse: Technology, Religion, And Ideology In Bryan Singer’S H+, Edward Brennan

Books/Book chapters

This essay critically analyses the digital series H+. In the near future, adults who can afford them, have replaced tablets and cell phones with nanotechnology implants. The H+ implant acts as a medical diagnostic and can overlay the user's senses with a computer interface. The apocalypse comes in the form of a computer virus which infects the H+ network and instantly kills one third of humanity. The series represents the anxiety and religiosity that surrounds the possible social consequences of digital technology. It also explores the tensions and intersections between technology and faith. This essay makes the case, however, that …

The Reconfigurable Machinery Efficient Workspace Analysis Based On The Twist Angles, Ana M. Djuric, Vukica Jovanovic, Mirjana Filipovic, Ljubinko Kevac

Engineering Technology Faculty Publications

A novel methodology for the calculation, visualisation and analysis of the Reconfigurable Machinery Efficient Workspace (RMEW), based on the twist angles, is presented in this paper. The machinery's kinematic parameters are used for calculating the workspace, while the efficient workspace is associated with the machinery's path and includes the end-effector position and orientation. To analyse and visualise many different machinery efficient workspaces at the same time, the calculation is based on the previously developed and validated complex reconfigurable machinery's kinematic structure named n-DOF Global Kinematic Model (n-GKM). An industrial robot is used as an example to demonstrate …

Integrating Cobots In Engineering Technology Education, Ana M. Djuric, Vukica Jovanovic, Tatiana V. Goris, Otilia Popescu

Engineering Technology Faculty Publications

Collaborative robots or CoBots, unlike traditional robots, are safe and flexible enough to work harmoniously with humans. Exploiting the efficiency of automated operations and the flexibility of manual operations in one process can improve productivity and worker job satisfaction. CoBots technology has been experiencing strong growth in different areas such as ground transportation, food-processing industry, car manufacturing, and naval or aeronautical engineering. Current CoBots education and training opportunities are rare or non-existent in university environments. In response to this need, we developed several CoBots modules which will be integrated in the current robotics and mechatronics courses. In this paper we …

An Improved Smote Algorithm Based On Genetic Algorithm For Imbalanced Data Collection, Qiong Gu, Xian-Ming Wang, Zhao Wu, Bing Ning, Chun-Sheng Xin

Electrical & Computer Engineering Faculty Publications

Classification of imbalanced data has been recognized as a crucial problem in machine learning and data mining. In an imbalanced dataset, minority class instances are likely to be misclassified. When the synthetic minority over-sampling technique (SMOTE) is applied in imbalanced dataset classification, the same sampling rate is set for all samples of the minority class in the process of synthesizing new samples, this scenario involves blindness. To overcome this problem, an improved SMOTE algorithm based on genetic algorithm (GA), namely, GASMOTE was proposed. First, GASMOTE set different sampling rates for different minority class samples. A combination of the sampling rates …

Automated Design Of Boolean Satisfiability Solvers Employing Evolutionary Computation, Alex Raymond Bertels

Masters Theses

"Modern society gives rise to complex problems which sometimes lend themselves to being transformed into Boolean satisfiability (SAT) decision problems; this thesis presents an example from the program understanding domain. Current conflict-driven clause learning (CDCL) SAT solvers employ all-purpose heuristics for making decisions when finding truth assignments for arbitrary logical expressions called SAT instances. The instances derived from a particular problem class exhibit a unique underlying structure which impacts a solver's effectiveness. Thus, tailoring the solver heuristics to a particular problem class can significantly enhance the solver's performance; however, manual specialization is very labor intensive. Automated development may apply hyper-heuristics …

Cp-Nets: From Theory To Practice, Thomas E. Allen

Theses and Dissertations--Computer Science

Conditional preference networks (CP-nets) exploit the power of ceteris paribus rules to represent preferences over combinatorial decision domains compactly. CP-nets have much appeal. However, their study has not yet advanced sufficiently for their widespread use in real-world applications. Known algorithms for deciding dominance---whether one outcome is better than another with respect to a CP-net---require exponential time. Data for CP-nets are difficult to obtain: human subjects data over combinatorial domains are not readily available, and earlier work on random generation is also problematic. Also, much of the research on CP-nets makes strong, often unrealistic assumptions, such as that decision variables must …

Modeling, Learning And Reasoning About Preference Trees Over Combinatorial Domains, Xudong Liu

Theses and Dissertations--Computer Science

In my Ph.D. dissertation, I have studied problems arising in various aspects of preferences: preference modeling, preference learning, and preference reasoning, when preferences concern outcomes ranging over combinatorial domains. Preferences is a major research component in artificial intelligence (AI) and decision theory, and is closely related to the social choice theory considered by economists and political scientists. In my dissertation, I have exploited emerging connections between preferences in AI and social choice theory. Most of my research is on qualitative preference representations that extend and combine existing formalisms such as conditional preference nets, lexicographic preference trees, answer-set optimization programs, possibilistic …

Preferences: Optimization, Importance Learning And Strategic Behaviors, Ying Zhu

Theses and Dissertations--Computer Science

Preferences are fundamental to decision making and play an important role in artificial intelligence. Our research focuses on three group of problems based on the preference formalism Answer Set Optimization (ASO): preference aggregation problems such as computing optimal (near optimal) solutions, strategic behaviors in preference representation, and learning ranks (weights) for preferences.

In the first group of problems, of interest are optimal outcomes, that is, outcomes that are optimal with respect to the preorder defined by the preference rules. In this work, we consider computational problems concerning optimal outcomes. We propose, implement and study methods to compute an optimal outcome; …

Universal Memory Architectures For Autonomous Machines, Dan Guralnik, Daniel E. Koditschek

Dan Guralnik

Iclp Tutorial: Relating Constraint Answer Set Programming And Satisfiability Modulo Theories, Yuliya Lierler

Yuliya Lierler

No abstract provided.

Systems, Engineering Environments, And Competitions, Yuliya Lierler, Marco Maratea, Francesco Ricca

Yuliya Lierler

Vascular Tree Structure: Fast Curvature Regularization And Validation, Egor Chesakov

Electronic Thesis and Dissertation Repository

This work addresses the challenging problem of accurate vessel structure analysis in high resolution 3D biomedical images. Typical segmentation methods fail on recent micro-CT data sets resolving near-capillary vessels due to limitations of standard first-order regularization models. While regularization is needed to address noise and partial volume issues in the data, we argue that extraction of thin tubular structures requires higher-order curvature-based regularization. There are no standard segmentation methods regularizing surface curvature in 3D that could be applied to large 3D volumes. However, we observe that standard measures for vessels structure are more concerned with topology, bifurcation angles, and other …

Pattern Discovery In Dna Using Stochastic Automata, Shweta Shweta

Master's Projects

We consider the problem of identifying similarities between different species of DNA. To do this we infer a stochastic finite automata from a given training data and compare it with a test data. The training and test data consist of DNA sequence of different species. Our method first identifies sentences in DNA. To identify sentences we read DNA sequence one character at a time, 3 characters form a codon and codons form proteins (also known as amino acid chains).Each amino acid in proteins belongs to a group. In total we have 5 groups’ polar, non-polar, acidic, basic and stop codons. …

Email Similarity Matching And Automatic Reply Generation Using Statistical Topic Modeling And Machine Learning, Zachery L. Schiller

Electronic Theses and Dissertations

Responding to email is a time-consuming task that is a requirement for most professions. Many people find themselves answering the same questions over and over, repeatedly replying with answers they have written previously either in whole or in part. In this thesis, the Automatic Mail Reply (AMR) system is implemented to help with repeated email response creation. The system uses past email interactions and, through unsupervised statistical learning, attempts to recover relevant information to give to the user to assist in writing their reply.

Three statistical learning models, term frequency-inverse document frequency (tf-idf), Latent Semantic Analysis (LSA), and Latent Dirichlet …

Applying Bayesian Machine Learning Methods To Theoretical Surface Science, Shane Carr

McKelvey School of Engineering Theses & Dissertations

Machine learning is a rapidly evolving field in computer science with increasingly many applications to other domains. In this thesis, I present a Bayesian machine learning approach to solving a problem in theoretical surface science: calculating the preferred active site on a catalyst surface for a given adsorbate molecule. I formulate the problem as a low-dimensional objective function. I show how the objective function can be approximated into a certain confidence interval using just one iteration of the self-consistent field (SCF) loop in density functional theory (DFT). I then use Bayesian optimization to perform a global search for the solution. …

Detection Of Diabetic Foot Ulcers Using Svm Based Classification, Lei Wang, Peder Pedersen, Diane Strong, Bengisu Tulu, Emmanuel Agu, Qian He, Ronald Ignotz, Raymond Dunn, David Harlan, Sherry Pagoto

Emmanuel O. Agu

Diabetic foot ulcers represent a significant health issue, for both patients’ quality of life and healthcare system costs. Currently, wound care is mainly based on visual assessment of wound size, which suffers from lack of accuracy and consistency. Hence, a more quantitative and computer-based method is needed. Supervised machine learning based object recognition is an attractive option, using training sample images with boundaries labeled by experienced clinicians. We use forty sample images collected from the UMASS Wound Clinic by tracking 8 subjects over 6 months with a smartphone camera. To maintain a consistent imaging environment and facilitate the capture process …

Robust Distributed Scheduling Via Time Period Aggregation, Shih-Fen Cheng, John Tajan, Hoong Chuin Lau

Shih-Fen Cheng

In this paper, we evaluate whether the robustness of a market mechanism that allocates complementary resources could be improved through the aggregation of time periods in which resources are consumed. In particular, we study a multi-round combinatorial auction that is built on a general equilibrium framework. We adopt the general equilibrium framework and the particular combinatorial auction design from the literature, and we investigate the benefits and the limitation of time-period aggregation when demand-side uncertainties are introduced. By using simulation experiments on a real-life resource allocation problem from a container port, we show that, under stochastic conditions, the performance variation …

Robust Distributed Scheduling Via Time Period Aggregation, Shih-Fen Cheng, John Tajan, Hoong Chuin Lau

Shih-Fen Cheng

In this paper, we evaluate whether the robustness of a market mechanism that allocates complementary resources could be improved through the aggregation of time periods in which resources are consumed. In particular, we study a multi-round combinatorial auction that is built on a general equilibrium framework. We adopt the general equilibrium framework and the particular combinatorial auction design from the literature, and we investigate the benefits and the limitation of time-period aggregation when demand-side uncertainties are introduced. By using simulation experiments on a real-life resource allocation problem from a container port, we show that, under stochastic conditions, the performance variation …

Robust Distributed Scheduling Via Time Period Aggregation, Shih-Fen Cheng, John Tajan, Hoong Chuin Lau

Shih-Fen CHENG

In this paper, we evaluate whether the robustness of a market mechanism that allocates complementary resources could be improved through the aggregation of time periods in which resources are consumed. In particular, we study a multi-round combinatorial auction that is built on a general equilibrium framework. We adopt the general equilibrium framework and the particular combinatorial auction design from the literature, and we investigate the benefits and the limitation of time-period aggregation when demand-side uncertainties are introduced. By using simulation experiments on a real-life resource allocation problem from a container port, we show that, under stochastic conditions, the performance variation …

Predicting Energy Demand Peak Using M5 Model Trees, Sara S. Abdelkader, Katarina Grolinger, Miriam Am Capretz

Electrical and Computer Engineering Publications

Predicting energy demand peak is a key factor for reducing energy demand and electricity bills for commercial customers. Features influencing energy demand are many and complex, such as occupant behaviours and temperature. Feature selection can decrease prediction model complexity without sacrificing performance. In this paper, features were selected based on their multiple linear regression correlation coefficients. This paper discusses the capabilities of M5 model trees in energy demand prediction for commercial buildings. M5 model trees are similar to regression trees; however they are more suitable for continuous prediction problems. The M5 model tree prediction was developed based on a selected …