Student Theses Supervised

  • Name: Zhen Li
    Date: PhD, UGA, Summer 2013
    Title: Programming with Concurrency: Barriers to Learning and Explorations in Teaching
    Abstract: Programming activities are not trivial tasks. Rather, to carry out such complex problem solving tasks, programming expertise developed through long-term experience is required. In contrast, the current landscape of computing demands more reliable and efficient implementations of concurrent software programs. In this thesis work, we research topics in psychology of programming and computer science education with an emphasis on programming with concurrency to inform the fields of psychology of programming, computer science education, empirical software engineering and a broader scope of human factors related fields. We identify the barriers to learning programming with concurrency through review and empirical work. We synthesize the previous research findings with regard to programming expertise, generalize a conceptual framework of the development and application of programming expertise and indicate the importance of the knowledge repository component. We reveal the structure of concurrency-related concepts, and provide insight into the acquisition procedures for such knowledge with our description of a “misconception hierarchy” grounded from qualitative analysis of empirical data. Comprehensive arguments generated through a case study are further provided to describe non-concurrency-related barriers that are critical for students to learn and appreciate programming with concurrency.
  • Name: Varuna Iyengar
    Date: MS, UGA, December 2012
    Title: MedAdherence: Cognitive Behavioral Therapy Via Mobile Web for Improved Medication Adherence
    Abstract: After an organ transplant, adherence to a prescribed medication regimen, is vital after-care to avoid organ rejection and ultimately, death. For adolescents, barriers to medication adherence are often tied to social, emotional, and behavioral issues rather than simple forgetfulness. Studies have identified Cognitive Behavioral Therapy (CBT) as an effective approach for improving medication adherence in adolescents. However, following up with patients after every dose and helping them overcome their barriersto is not a feasible solution.MedAdherence is aSmartphone-based CBT system that incorporates incentives and is designed to improve medication adherence in adolescent kidney transplant patients. The system sends reminders, records adherence data, and provides support for CBT at the moment of non-adherence. In this work, we also evaluate the feasibility of the LogicBlox® platform, adatalog-based, Cloud-delivered platform-as-a-service, formobile web applications such as MedAdherence.
  • Name: Philippa Rhodes
    Date: PhD, UGA, Summer 2007
    Title: Software Visualization: Using Perceptual, Attentional, and Cognitive Concepts to Quantify Quality and Improve Effectiveness
    Abstract: Software visualization (SV) involves the use of the crafts of typography, graphic design, animation, and cinematography with modern Human-Computer Interaction technology to facilitate both the human understanding and effective use of computer software. Software visualizations are often used to portray both concrete and abstract concepts and range from depictions of source code to performance characteristics to the execution of an algorithm as a discrete or continuous sequence of graphical images, or algorithm visualization. Numerous algorithm visualizations have been developed for use in educational settings. However, studies that were designed to demonstrate the pedagogic effectiveness of algorithm visualizations have been markedly unsuccessful, in spite of high expectations. In response to these results, lists of recommended features have been suggested to algorithm visualization system designers, but most of these features have not been proven to be beneficial. The broad goal of this research is to provide an empirically-validated method for designing and evaluating the effectiveness of dynamic visualizations. Our approach has been to identify features of these visualizations and systems that may improve learning, to create software that can isolate features of interest and aid in evaluating the usefulness of these features, and to then use the software to conduct and analyze user studies. This research: i) assembles an initial listing of features of SVs and SV systems and introduces a framework for testing the effectiveness of each, ii) provides verified design guidelines for dynamic visualizations, iii) applies concepts already researched and established in perceptual psychology and cognitive psychology to the design of effective SVs, iv) offers an explanation of the inevitable variability present in studies involving human subjects through the investigation of the effects of individual differences on comprehending SVs, and v) objectively classifies SV systems and makes them widely and easily available in a way never done before, with VisIOn, an Interactive Visualization Ontology.
  • Name: Manish Agarwal
    Date: MS, UGA, Spring 2007
    Title: Viewing Behavior Model Graphs (VBMGs) for Characterizing User Viewing Behavior in Program Visualizations
    Abstract: We present a methodology for characterizing the gaze behavior of viewers of animated displays. We introduce a transition graph called Viewing Behavior Model Graph (VBMG) to characterize the behavior of users with similar viewing patterns into separate groups. We apply this methodology to the viewing of program visualizations. In this method the user's eye-fixation sequences are obtained using an Eye-Tracker and per-user viewing behavior models are created. We then cluster these per-user models to build VBMGs for each cluster. The VBMGs are useful because they help us classify users into separate groups, each user within a group having viewing behavior similar to others in the group. One useful application of VBMGs would be to dynamically capture viewing behavior and predict the cluster to which a user belongs, thus permitting on-the-fly adaptation of displays and other teaching materials.
  • Name: Shradha Kaldate
    Date: MS, UGA, Spring 2007
    Title: Analysis of Viewing Behavior of Program Visualization and Interaction with Individual Differences
    Abstract: Program Visualization (PV) is believed to be useful in Computer Science education. However, while some PVs have been found to help users learn, other PVs have not been beneficial. In this thesis we studied the user's gaze pattern to find effects of popup questions on an individual's visual attention. We further analyzed the correlation of gaze behavior with the individual's preferred learning style, performance based on a pre-test and post-test and a variety of perceptual, attentional and cognitive abilities as determined by a battery of paper-and-pencil and computer-based assessments. While popups appear to be effective for directing attention, no significant effect on comprehension of the depicted algorithm was detected. Individual with different learning styles have distinctive viewing patterns and this finding should be utilized in designing PVs that consider individual differences.
  • Name: Hongyu Yang
    Date: MS, UGA, Spring 2007
    Title: Phylogenetic Tree Display: A Web-based Visualization Tool of Phylogenetic Data
    Abstract: To be added.
  • Name: Ashley George Hamilton-Taylor
    Date: PhD, UGA, Summer 2006
    Title: The Study and Design of Algorithm Animations
    Abstract: Algorithm Animations (AAs) portray the high-level dynamic operation of an algorithm. The computer science education community held great expectations that AA would assist students in understanding algorithms. However, many studies of the instructional effectiveness of AAs have produced inconclusive results. We investigated a number of issues pertinent to AA effectiveness: the study of AA user needs, user-centered design and the role of perception in AA. Existing algorithm animation systems typically have been designed without formal study of related teaching practices. We conducted an observational study of instructors teaching data structure and algorithm topics, focusing on activities involving the use of diagrams and algorithms. The results of this study were used to inform the user-centered design of SKA, the Support Kit for Animation. SKA combines interactive data structure diagram manipulation with flexible pseudocode execution, simple algorithm animation authoring support, a visual data structure library, and an animation engine designed for perceptual pacing and timing. The role of perception in AAs had not been formally considered in the past. We collaborated on a number of empirical studies to investigate this role, and the design of software to be used to conduct these studies. We found that some animation techniques can assist user perception and mapping in AA in some contexts, which will inform future AA design and studies.
  • Name: Bina Reed
    Date: MS, UGA, Spring 2006
    Title: Investigating characteristics of effective program visualizations: a testing environment and the effect of comparison cueing and exchange techniques on viewer comprehension in algorithm animations
    Abstract: We describe SSEA, a System for Studying the Effectiveness of Animations, and an empirical study that evaluates two visualization design attributes. SSEA was created as a testing environment for studying the effects of various attributes in visualization design on viewer comprehension. Researchers create a series of animations in SSEA with a design characteristic in mind. SSEA allows these animations to be viewed while recording the viewer's interactions and responses to questions about the underlying algorithm. At the conclusion of running all experiments, the researchers can examine the log files generated, and perform analysis of the responses and timings with respect to the attribute being examined. The first in a series of studies to be conducted examined the attributes: comparison queueing and exchange techniqe using traditional and popup questions, measuring comprehension and perception, respectively. No significant effects were observed in comprehension questions. Significant effects were found in the perceptual questions.
  • Name: Nivedita Kaluskar
    Date: MS, UGA, Fall 2005
    Title: XML-based specification and automatic code generation for easy customization of view in the GUS WDK framework
    Abstract: An automatic custom view generation method for the Genomics Unified Schema (GUS) frame-work has been designed and implemented in this thesis work. The Genomics Unified Schema is a database schema and application framework used to store, integrate, analyze and present functional genomics data. Cryptosporidium, The Genome Resource (CryptoDB) is a project developed at the University of Georgia under the GUS framework. CryptoDB offers a number of queries that a user can ask of the database and uses the GUS Web Development Kit (WDK) to generate a default view for the database. Although the earlier WDK offered an easy way of generatinga view for any given data model, the view generated was too closely tied to Model and Controller. The method for generating a custom view enables a more clear separation between the Model, View and Controller in the MVC type application and allows on-the-fly generation of custom views. It uses an XML-based custom view specifications file containing the specifications a user is allowed to make in order to generate a custom view. This gives the user much more flexibility in designing a custom view and allows the View to be less closely tied to the Model and the Controller.
  • Name: Yanqi Su
    Date: MS, UGA, Fall 2005
    Title: Comparative genomics visualization
    Abstract: As an increasing number of genomes are sequenced, comparative genomics analysis at different evolutionary distances plays a crucial role in decoding genomic information and discovering the similarities and differences between the genomes. It consists of three process steps: data preparation, sequence alignment, and visualization. In this thesis the major alignment methods and visualization tools are discussed. Based on an analysis of current work, we have designed and implemented a comparison visualization approach that satisfies the specific requirements of the CryptoDB project (http://cryptodb.org). Our approach is written in Perl and based on the GBrowse framework. In addition, this study also presents an analysis of the comparison of C. hominis and C. parvum to find candidate insertions, deletions, and synteny blocks. Future work is also discussed.
  • Name: Sujith Thomas
    Date: MS, UGA, Fall 2004
    Title: An experiment designer tool for evaluation of program visualization quality
    Abstract: Program Visualization refers to the graphical representation of a program in execution. They are used to facilitate better understanding of the underlying algorithm or other program behavior. However, doubts exist about the usefulness of program visualizations and the extent to which the visualizations are employed in practice. Some studies of program and algorithm visualizations have shown that visualizations have helped the instructional effort, while others argue that they have little or no effect on the learning of subjects. Explaining the discrepancy is part of the research to be carried out. One theory is that the quality of the program visualization used has varied among experiments. An experiment generation application called TestCreator has been developed to design experiments that assist in identification of the attributes associated with program visualization quality and quantification of the effects of these attributes on the ability of visualizations to facilitate learning.
  • Name: Matthew Ross
    Date: MS, UGA, Summer 2004
    Title: A testing environment for the evaluation of program visualization quality
    Abstract: Program visualizations have the potential to convey information about the behavior of the programs they depict. However, program visualizations in use vary widely in their ability to convey the desired information. In this work we conduct a study of the quality of the visualizations, the ability of those visualizations to present information in a way that is both effective and efficient in creating understanding. In order to measure program visualization quality, many traits of the visualization need to be considered: size, shape, color, and location of graphical elements, cueing and motion styles, etc. A testing environment called TestTaker has been developed that allows the experimenter to display visualizations to the user and receive feedback. The feedback is then used to evaluate the quality of the program visualizations based on the metrics specified. These quality metrics will be used to create a standard way of measuring the quality of program visualizations.
  • Name: Jian Wang
    Date: MS, UGA, Summer 2004
    Title: Feature search in biological sequence data
    Abstract: Genome projects continue to produce large quantities of sequence data. Annotation of this sequence data to indicate the location of genes, start and stop codons, inverted and direct repeats, and other patterns of interest is a challenging problem. In this thesis we present three contributions to solving this problem. First, we have performed an analysis of several gene-finding programs for the fungus N. crassa , and applied both standard metrics and new metrics we have defined. Next, we have developed a general tool that can automatically evaluate any gene-finding program and report performance metrics. Finally, we have developed an Interactive Pattern Search Tool (IPST) to facilitate finding complex patterns in nucleotide sequence data. The hashtable based approach employed in IPST is compared with the suffix tree approach for pattern search. IPST is applied to the problems of locating Long Terminal Repeat (LTR) retrotransposons and Miniature Inverted repeat Transposable Elements (MITEs) in rice sequences.
  • Name: Shiming Dong
    Date: MS, UGA, Summer 2004
    Title: Calculation, visualization, and manipulation of MASTs (Maximum Agreement Subtrees)
    Abstract: Phylogenetic trees are used to represent the evolutionary history of a set of species. Comparison of multiple phylogenetic trees can help researchers find the common classification of a tree group, compare tree construction inferences or obtain distances between trees. We present TreeAnalyzer, a freely available package for phylogenetic tree comparison. A MAST (Maximum Agreement Subtree) algorithm is implemented to compare the trees. Additional features of this software include tree comparison, visualization, manipulation, labeling, and printing.
  • Name: Vinay Sachdev (supervised jointly with Prof. Maria Hybinette)
    Date: MS, UGA, Spring 2004
    Title: Overcoming over-optimism in Time Warp via aggregation of fast processes
    Abstract: The parallel Time Warp protocol is widely used to improve the performance of large-scale discrete event simulations such as simulations of the air traffic control system and the World Wide Web. However, a concern regarding optimistic simulation processing is that some logical processes (LPs) may progress far beyond others into the simulated future, causing an imbalance that may degrade performance. The degradation is due to long and excessive rollbacks, inefficient use of memory resources, and communication overheads. Allowing this "over-optimistic" processing may bring the simulation to a halt. This is unacceptable for long running simulations that may take days to complete. We present a new mechanism that controls over-optimistic processing through aggregation and isolation of fast processes and redistribution of the slow, less optimistic processes. Our performance results demonstrate that our techniques can improve the useful work by a factor of 1.75 and also improve execution time, while keeping the overhead small.
  • Name: Qin Zhang
    Date: MS, UGA, Spring 2004
    Title: Mini-Lab: A Tool to Visualize Normal Modes of Vibration Using Java-3D
    Abstract: Vibrational analysis of polyatomic molecules requires tools with the capability to construct the system and display the normal modes. Mini-Lab, developed in this thesis work, is a 3D visualization tool and vibration analyzer that serves both educational and research needs. The goal is to help the user visualize the atoms and bonds, compute the normal modes for the system, and generate the animated vibrational pattern. Users can experiment with different masses and bonds to construct a vibration system and understand the normal modes when studying harmonic oscillation or small vibrations in classical mechanics. This tool has been used in research that analyzes the vibration of nanocrystals.
  • Name: Weicheng Zhang
    Date: MS, UGA, Summer 2003
    Title: A user friendly environment for gene-finding program evaluation (GFPE)
    Abstract: A graphical, user friendly environment for GFPE (Gene-Finding Program Evaluation), written in Java, is developed to evaluate the prediction accuracy of gene-finding programs GenScan, HMMGene, GeneMark, Pombe and FFG. This tool aims to simplify and/or automate the process of executing the gene-finding programs on sequences of interest and of collecting and analyzing the results. The GUI is designed to be similar to a spreadsheet table. The user can add, cut, copy or paste gene sequence files and annotation files to the table, add, delete, or modify the GFPE program and parameters, select an area on the table to represent the execution of various gene-finding programs on remote servers, automatically collect the prediction results, and draw bar charts to compare evaluation accuracy at the coding level, exon level and protein level. It provides a convenient, user-friendly environment and an efficient file management and execution system, thus saving the user substantial time.
  • Name: Chetna Warade
    Date: MS, UGA, Fall 2003
    Title: Web Services Composition for Microarray Data Analysis
    Abstract: To be added.
  • Name: Arumugaraja Selvaraj
    Date: MS, UGA, Fall 2002
    Title: Interactive computational steering: conservative vs optimistic steering approaches
    Abstract: Interactive Computational Steering is the online, interactive allocation and adjustment of system resources and application parameters. Causal consistency is an important feature of interactive steering of distributed computations, as it is often required to maintain the correctness of the computation. However, due to the asynchronous nature of distributed computations, it is difficult to coordinate steering changes across processes to guarantee that the changes are applied consistently at all processes. Two general approaches exist for achieving interactive computational steering: conservative steering and optimistic steering. In this thesis, we present algorithms for the conservative steering approach. We also compare the performance of the conservative steering approach with that of the optimistic steering approach with regard to perturbation and lag.
  • Name: Jinhua Guo
    Date: PhD, UGA, Summer 2002
    Title: Consistent, Interactive Steering of Distributed Computations: Algorithms and Implementation
    Abstract: Interactive computational steering provides users with the opportunity to tackle new problems in a way that helps them to learn about the computation in a highly engaging, interactive, visual environment. Causal consistency is an important feature of interactive steering of distributed computations, as it is often required to maintain the correctness of the computation. However, due to the asynchronous nature of distributed computations, it is difficult to coordinate steering changes across processes to guarantee that the changes are applied consistently at all processes. This thesis introduces a transaction-based computation model for distributed computation. This abstract model not only gives users a simple and high-level view of distributed computation, but also simplifies reasoning about the consistency problem by reducing the amount of information to be handled. Furthermore, this work investigates two approaches for achieving consistent steering: conservative steering and optimistic steering. The performance of conservative and optimistic steering approaches is evaluated in terms of perturbation and lag. Our experiments show that when the percentages of consistency on the first attempt are large enough and the size of checkpoint is not too large, the optimistic approach will achieve better performance; otherwise, the conservative approach will be better.
  • Name: Mihail Tudoreanu (supervised jointly with Prof. Gruia-Catalin Roman)
    Date: DSc, WUSTL, Spring 2002
    Title: Economy of Interaction in Program Visualization: Designing Effective Visualization Tools for Reducing User's Cognitive Effort
    Abstract: Program visualization has the potential to be an important tool for people who seek to observe and understand the behavior of a running computation. This thesis focuses on alleviating barriers to the realization of this potential that pertain to the design of a visualization system and to insufficient knowledge about how people take advantage of program visualizations. Our major contribution is the design of a visualization approach capable of improving user's performance through the use of economy of information and tasks. We present evidence from our empirical studies that this type of economy promotes animations capable of significantly improving people's understanding of the computation. We apply this knowledge to develop a system for creating application-specific visualizations solely through interactions with program visualizations and textual views of the computation, thus promoting economy of interaction. The system is built around the principle that animation viewers are also the creators of animations and systematically refine the visualizations to suit their momentary goal.
  • Name: David Miller
    Date: MS, UGA, Spring 2002
    Title: An optimistic approach to computational steering
    Abstract: Computational Steering is the online, interactive allocation of resources and adjustment of application parameters. Few computational steering systems support the coordinated steering of multiple processes. Of those that do provide such support, our system is unique in its optimistic approach; other systems take a conservative approach. Because this requires global synchronization considerable perturbation can arise. We focus on optimistic steering, which does not require global synchronization before a steering event may take place. To achieve this requires not only the ability to determine the consistency of steering transactions but also the ability to correct any inconsistencies that may occur. To address these issues, we have developed algorithms for consistency detection and a steering system that has the ability to correct inconsistencies through computational rollback and re-execution. Presented in this thesis are both the details of our steering system and a performance analysis of that system.
  • Name: Tao Wu
    Date: MS, UGA, Fall 2001
    Title: An extensible framework for developing visualization software for gene expression data
    Abstract: A design architecture of a software system that analyzes and visualizes gene expression profiles is proposed. Implementation of core elements in the system is provided and a functional system prototype is developed. This system is capable of visualizing data through a scrollable user interface (UI) that only uses a small amount of memory. Current system includes implementation for viewing data as a rooted tree or colored heat maps and includes support for clustering, filtering and printing.
  • Name: Ritu Dhawan
    Date: MS, UGA, Fall 2001
    Title: Evaluation of web personalization software and visualization with the help of usability study
    Abstract: Evaluation is an important element in the creation of complex software. In this thesis, we describe two such evaluations. In the first study, we perform usability testing on web-based personalization software developed by John O'Looney of the Carl Vinson Institute of Government at the University of Georgia. Personalization of a web page involves dynamically altering the contents of the web pages according to the preferences of customers or clients, so that each client gets information specific to his needs. This study assists in and suggests ways to improve the usability and learnability of this software. The second study was a user study to determine if visualization is beneficial in helping users to learn about and understand distributed algorithms.
  • Name: Rong Wu
    Date: MS, UGA, Fall 2001
    Title: Visualization as an aid for understanding distributed algorithms
    Abstract: Visualization, the graphical representation of data by computer, is believed to be an aid to cognitive processes. Algorithm animation, a type of visualization that displays a dynamic visual representation of the state and behavior of algorithms, is aimed to help students learn and understand complex algorithms better and more easily. Previous evaluations of algorithm animations have been performed. However, the results of these studies were mixed and often inconclusive. The visualization community has called for further studies to answer the question of whether algorithm animation is beneficial. This thesis presents such a study. Three-dimensional graphics were used to implement algorithm animations for distributed algorithms. Two studies were conducted to evaluate whether these visualizations helped students to learn about the distributed algorithms.
  • Name: Brandon Lee Kohn
    Date: MS, UGA, Summer 2001
    Title: Practical considerations in monitoring and steering of distributed computations
    Abstract: Tools that assist users in understanding the complex workings of distributed systems are an important aid to programmers and scientists. Interactive monitoring and steering tools can provide users with insight into their computations. This thesis describes the PathFinder interactive monitoring and steering system and the problems involved in making the system a viable solution for monitoring and steering distributed applications. These problems include MPI support, an Agent Wizard, and a performance analysis of two agent subsystems and their associated milieu, one implemented in Perl and the other in Java.
  • Name: Renyi Liu
    Date: MS, UGA, Summer 2001
    Title: Strategies for Improving Multiple Alignment of Retrotransposon Sequences
    Abstract: Multiple sequence alignment plays a crucial role in extracting structural, functional, and evolutionary information from the exponentially growing sequence data from the ongoing genome sequencing. Based on the case study of retrotransposon sequence alignment, this thesis compares three alignment programs, DIALIGN, CLUSTALW, and PRRN, and proposes some strategies to improve alignment quality, such as realigning certain sequences or sequence ranges with different programs or parameters and hand editing. Entropy is used as an alignment quality indicator. This study also presents the design and development of an alignment tool, named AlignAgain, which is built to help biologists to improve alignment quality. AlignAgain is written in Java and allows users to display, edit, realign whole or partial sequences with CLUSTALW or PRRN, and append sequences with profile alignment.
  • Name: Piyush Burte
    Date: MS, UGA, Summer 2001
    Title: A Visualization-based Tool in Support of High Throughput Nuclear Magnetic Resonance Studies
    Abstract: To be added.
  • Name: Yin Xiong
    Date: MS, UGA, Summer 2001
    Title: An Exploratory Environment for Concurrency Control Algorithms
    Abstract: To be added.
  • Name: James Skinner
    Date: MAMS, UGA, Spring 2001
    Title: Deployment Interface Module for the United States Army
    Abstract: To be added.
  • Name: Yong Zhang
    Date: MS, UGA, Spring 2001
    Title: A Visualization System for Protein Interaction Mapping Using Java 3D Technology
    Abstract: To be added.
  • Name: Sneha Rao Kadandale
    Date: MS, UGA, Spring 2001
    Title: Visualizations in Support of Network Monitoring and Control
    Abstract: To be added.
  • Name: Navin Gupta
    Date: MS, UGA, Summer 2000
    Title: Performance Considerations in the Monitoring and Visualization of Distributed Computations
    Abstract: To be added.
  • Name: Delbert Hart (supervised jointly with Prof. Gruia-Catalin Roman)
    Date: DSc, WUSTL, Summer 2000
    Title: Supporting Exploratory Visualization of Distributed Computations
    Abstract: To be added.
  • Name: Himabindu Vuppula
    Date: MS, UGA, Spring 2000
    Title: Practical Algorithms for Snapshot Collection in Distributed Processes
    Abstract: The collection of global snapshots is an essential task in the monitoring and interactive steering of distributed systems. In the PathFinder exploratory visualization system we have implemented multiple snapshot collection algorithms, each providing a di erent balance among performance characteristics, especially consistency, lag, perturbation and scalability. In this thesis, we describe the PathFinder system, present the algorithms, describe experiments used to evaluate the algorithms, and present and explain the results of these experiments.


Clemson University School of Computing | Office: 104 McAdams Hall, Clemson, SC 29634 | 864.656.5874 | etkraem@clemson.edu