Theory and Algorithms Commons^™

Assessment Of Structure From Motion For Reconnaissance Augmentation And Bandwidth Usage Reduction, Jonathan B. Roeber

Theses and Dissertations

Modern militaries rely upon remote image sensors for real-time intelligence. A typical remote system consists of an unmanned aerial vehicle, or UAV, with an attached camera. A video stream is sent from the UAV, through a bandwidth-constrained satellite connection, to an intelligence processing unit. In this research, an upgrade to this method of collection is proposed. A set of synthetic images of a scene captured by a UAV in a virtual environment is sent to a pipeline of computer vision algorithms, collectively known as Structure from Motion. The output of Structure from Motion, a three-dimensional model, is then assessed in …

Cora: Commingled Remains And Analytics – An Open Community Ecosystem, Nicole Mcelroy, Ryan Ernst

UNO Student Research and Creative Activity Fair

Anthropologists at organizations such as the DPAA (Defense POW/MIA Accounting Agency) have the tough job of sorting through commingled remains of fallen soldiers. Under the direction of Professor Pawaskar at the College of IS&T, Ryan Ernst and I are currently developing a web application for the DPAA that will help them inventory the bones and record all the appropriate associations. After the inventory web application is built we will begin the analysis process using graph theory and other mathematical algorithms. This will ultimately help organizations like the DPAA get closer to the end goal of identifying fallen soldiers from commingled …

Virtualized Cloud Platform Management Using A Combined Neural Network And Wavelet Transform Strategy, Chunyu Liu

Electronic Theses, Projects, and Dissertations

This study focuses on implementing a log analysis strategy that combines a neural network algorithm and wavelet transform. Wavelet transform allows us to extract the important hidden information and features of the original time series log data and offers a precise framework for the analysis of input information. While neural network algorithm constitutes a powerfulnonlinear function approximation which can provide detection and prediction functions. The combination of the two techniques is based on the idea of using wavelet transform to denoise the log data by decomposing it into a set of coefficients, then feed the denoised data into a neural …

Target Detection Using Convolutional Neural Networks, Robert P. Loibl

Theses and Dissertations

This research explores the use of Convolutional Neural Networks (CNNs) to classify targets of interest within satellite imagery. Methods were specifically devised for the classification of airports within Landsat-8 scenes. A novel automated dataset generation technique was developed to create labeled datasets from satellite imagery using only coordinate metadata. Using this approach a very large dataset of over 132,000 labeled images was created without human input. This dataset was used to evaluate the effects of color and resolution on airport classification accuracy. Two experiments were run with the first experiment classifying large airports with 96.8% accuracy, and the second classifying …

Cryptosystems Using Subgroup Distortion, Indira Chatterji, Delaram Kahrobaei, Ni Yen Lu

Publications and Research

In this paper we propose cryptosystems based on subgroup distortion in hyperbolic groups. We also include concrete examples of hyperbolic groups as possible platforms.

Sparse Passive-Aggressive Learning For Bounded Online Kernel Methods, Jing Lu, Doyen Sahoo, Peilin Zhao, Steven C. H. Hoi

Research Collection School Of Computing and Information Systems

One critical deficiency of traditional online kernel learning methods is their unbounded and growing number of support vectors in the online learning process, making them inefficient and non-scalable for large-scale applications. Recent studies on scalable online kernel learning have attempted to overcome this shortcoming, e.g., by imposing a constant budget on the number of support vectors. Although they attempt to bound the number of support vectors at each online learning iteration, most of them fail to bound the number of support vectors for the final output hypothesis, which is often obtained by averaging the series of hypotheses over all the …

R3: Reinforced Ranker-Reader For Open-Domain Question Answering, Shuohang Wang, Mo Yu, Xiaoxiao Guo, Zhiguo Wang, Tim Klinger, Wei Zhang, Shiyu Chang, Gerald Tesauro, Bowen Zhou, Jing Jiang

Research Collection School Of Computing and Information Systems

In recent years researchers have achieved considerable success applying neural network methods to question answering (QA). These approaches have achieved state of the art results in simplified closed-domain settings such as the SQuAD (Rajpurkar et al. 2016) dataset, which provides a pre-selected passage, from which the answer to a given question may be extracted. More recently, researchers have begun to tackle open-domain QA, in which the model is given a question and access to a large corpus (e.g., wikipedia) instead of a pre-selected passage (Chen et al. 2017a). This setting is more complex as it requires large-scale search for relevant …

Cross-Language Learning For Program Classification Using Bilateral Tree-Based Convolutional Neural Networks, Duy Quoc Nghi Bui, Lingxiao Jiang, Yijun Yu

Research Collection School Of Computing and Information Systems

Towards the vision of translating code that implements an algorithm from one programming language into another, this paper proposes an approach for automated program classification using bilateral tree-based convolutional neural networks (BiTBCNNs). It is layered on top of two tree-based convolutional neural networks (TBCNNs), each of which recognizes the algorithm of code written in an individual programming language. The combination layer of the networks recognizes the similarities and differences among code in different programming languages. The BiTBCNNs are trained using the source code in different languages but known to implement the same algorithms and/or functionalities. For a preliminary evaluation, we …

Relating Justification Logic Modality And Type Theory In Curry–Howard Fashion, Konstantinos Pouliasis

Dissertations, Theses, and Capstone Projects

This dissertation is a work in the intersection of Justification Logic and Curry--Howard Isomorphism. Justification logic is an umbrella of modal logics of knowledge with explicit evidence. Justification logics have been used to tackle traditional problems in proof theory (in relation to Godel's provability) and philosophy (Gettier examples, Russel's barn paradox). The Curry--Howard Isomorphism or proofs-as-programs is an understanding of logic that places logical studies in conjunction with type theory and -- in current developments -- category theory. The point being that understanding a system as a logic, a typed calculus and, a language of a class of categories constitutes …

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 …

Two Birds With One Stone: Classifying Positive And Unlabeled Examples On Uncertain Data Streams, Donghong Han, Shuoru Li, Fulin Wei, Yuying Tang, Feida Zhu, Guoren Wang

Research Collection School Of Computing and Information Systems

An important feature characteristic of the data streams in many of today's big data applications is the intrinsic uncertainty, which could happen for both item occurrence and attribute value. While this has already posed great challenges for fundamental data mining tasks such as classification, things are made even more complicated by the fact that completely-labeled examples are usually unavailable in such settings, leaving researchers the only option to learn classifiers on partially-labeled examples on uncertain data streams. Furthermore, there will be concept drift on evolving data streams. To address these challenges, this paper therefore focuses on the study of learning …

Models As Weapons: Review Of Weapons Of Math Destruction: How Big Data Increases Inequality And Threatens Democracy By Cathy O’Neil (2016), Samuel L. Tunstall

Numeracy

Cathy O’Neil. 2016. Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy (New York, NY: Crown) 272 pp. ISBN 978-0553418811.

Accessible to a wide readership, Cathy O’Neil’s Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy provides a lucid yet alarming account of the extensive reach of mathematical models in influencing all of our lives. With a particular eye towards social justice, O’Neil not only warns modelers to be cognizant of the effects of their work on real people—especially vulnerable groups who have less power to fight back—but also encourages laypersons to take initiative …

A Practical And Efficient Algorithm For The K-Mismatch Shortest Unique Substring Finding Problem, Daniel Robert Allen

EWU Masters Thesis Collection

This thesis revisits the k-mismatch shortest unique substring (SUS) finding problem and demonstrates that a technique recently presented in the context of solving the k-mismatch average common substring problem can be adapted and combined with parts of the existing solution, resulting in a new algorithm which has expected time complexity of O(n log^k n), while maintaining a practical space complexity at O(kn), where n is the string length. When k > 0, which is the hard case, the new proposal significantly improves the any-case O(n²) time complexity of the prior best method for k-mismatch SUS finding. Experimental study …

Gpu Accelerated Risk Quantification, Forrest L. Ireland

EWU Masters Thesis Collection

Factor Analysis of Information Risk (FAIR) is a standard model for quantitatively estimating cybersecurity risks and has been implemented as a sequential Monte Carlo simulation in the RiskLens and FAIR-U applications. Monte Carlo simulations employ random sampling techniques to model certain systems through the course of many iterations. Due to their sequential nature, FAIR simulations in these applications are limited in the number of iterations they can perform in a reasonable amount of time. One method that has been extensively used to speed up Monte Carlo simulations is to implement them to take advantage of the massive parallelization available when …

Glyph Based Segmentation Of Chinese Calligraphy Characters In The "Collected Characters" Stele., David A. Mcinnis

EWU Masters Thesis Collection

Text character segmentation is the process of detecting the bounding box position of individual characters within a written text document image. The character segmentation problem remains extremely difficult for ancient Chinese calligraphy documents. This paper examines a glyph-based segmentation technique for segmenting Chinese Calligraphy characters in the "Collected Characters". The glyph-based character segmentation pipeline utilizes a combination of well-understood image processing techniques in a novel pipeline which is able to detect Chinese calligraphy characters from ink-blots with a good reliability.

Mechanism Design, Matching Theory And The Stable Roommates Problem, Yashaswi Mohanty

Honors Theses

This thesis consists of two independent albeit related chapters. The first chapter introduces concepts from mechanism design and matching theory, and discusses potential applications of this theory, particularly in relation to dorm allocations in colleges. The second chapter investigates a subset of the dorm allocation problem, namely that of matching roommates. In particular, the paper looks at the probability of solvability of random instances of the stable roommates game under the condition that preferences are not completely random and exogenous but endogenously determined through a dependence on room choice. These probabilities are estimated using Monte-Carlo simulations and then compared with …

Efficient Randomized Algorithms For The Fixed Precision Low Rank Matrix Approximation, Wenjian Yu, Yu Gu, Yaohang Li

Computer Science Faculty Publications

Randomized algorithms for low-rank matrix approximation are investigated, with the emphasis on the fixed-precision problem and computational efficiency for handling large matrices. The algorithms are based on the so-called QB factorization, where Q is an orthonormal matrix. First, a mechanism for calculating the approximation error in the Frobenius norm is proposed, which enables efficient adaptive rank determination for a large and/or sparse matrix. It can be combined with any QB-form factorization algorithm in which B's rows are incrementally generated. Based on the blocked randQB algorithm by Martinsson and Voronin, this results in an algorithm called randQB_EI. Then, we further revise …

Code: Coherence Based Decision Boundaries For Feature Correspondence, Wen-Yan Lin, Fan Wang, Ming-Ming Cheng, Sai-Kit Yeung, Philip H. S. Torr, Jiangbo Lu

Research Collection School Of Computing and Information Systems

A key challenge in feature correspondence is the difficulty in differentiating true and false matches at a local descriptor level. This forces adoption of strict similarity thresholds that discard many true matches. However, if analyzed at a global level, false matches are usually randomly scattered while true matches tend to be coherent (clustered around a few dominant motions), thus creating a coherence based separability constraint. This paper proposes a non-linear regression technique that can discover such a coherence based separability constraint from highly noisy matches and embed it into a correspondence likelihood model. Once computed, the model can filter the …

Extensions Of The Morse-Hedlund Theorem, Eben Blaisdell

Honors Theses

Bi-infinite words are sequences of characters that are infinite forwards and backwards; for example "...ababababab...". The Morse-Hedlund theorem says that a bi-infinite word f repeats itself, in at most n letters, if and only if the number of distinct subwords of length n is at most n. Using the example, "...ababababab...", there are 2 subwords of length 3, namely "aba" and "bab". Since 2 is less than 3, we must have that "...ababababab..." repeats itself after at most 3 letters. In fact it does repeat itself every two letters. …

Bi-Objective Optimization Of Kidney Exchanges, Siyao Xu

Theses and Dissertations--Computer Science

Matching people to their preferences is an algorithmic topic with real world applications. One such application is the kidney exchange. The best "cure" for patients whose kidneys are failing is to replace it with a healthy one. Unfortunately, biological factors (e.g., blood type) constrain the number of possible replacements. Kidney exchanges seek to alleviate some of this pressure by allowing donors to give their kidney to a patient besides the one they most care about and in turn the donor for that patient gives her kidney to the patient that this first donor most cares about. Roth et al.~first discussed …