Platinum Plus Supporter
SIGCSE 2013 - Student Research Competition
Preserving Data Privacy Through Data Partitioning in Mobile Application
Mohammad Al-Mutawa, University of Colorado
To extend the capabilities of mobile devices, one can off-load computing tasks to more powerful remote nodes, which in many cases can improve performance and/or reduce power consumption. However, one of the problems associated with off-loading is the lose of data privacy. Our work use the concept of Data Partitioning to minimize, and in some case eliminate, the lose of personal data privacy. The overall execution consists of identifying sensitive data, shipping code and non-private data for remote execution and getting the results back, and then combining the results from local and remote executions on the mobile device.
Green Dolphin: a question and answer site for novice programmers
Chulakorn Aritajati, Auburn University
The Green Dolphin (GD) is a question and answer website. GD is designed to fulfill students’ need for collaborative learning by allowing them to interact with and ask questions of the professional community consisting of their professors, TAs and classmates. GD gives students automatic feedbacks of their code quality. GD provides an expert system to suggest students who are experts. Therefore, users get fast and high quality answers, and professors spend less time to communicate with students. Students can gain new knowledge from the flow of questions and answers in the system. Students have opportunities to learn communication skills as well as programming and debugging skills.
Wacky Writing: Enhancing the XO Laptop Platform to Motivate Creative Writing by Children
Austin C. Bart, Virginia Tech
Creative Writing is an important expression of creativity, and there currently exists no satisfactory Learning Software for the widely-distributed XO Laptop Platform to fill this niche. This study created two new Creative Writing Learning Activities for the XO Laptop and intervened in a classroom to test their effect on the Intrinsic Motivation of children to write creatively using the XO laptop. The quantitative and qualitative results indicate that children were motivated to write using the software, leading the way to further improvements to the software for future benefit.
SLASH: Scratch-based visual programming in Second Life for introductory computer science education
Kara A. Behnke, ATLAS Institute
This research describes the application of SLASH, a modification of Scratch that generates syntax in Linden Scripting Language (LSL) through block-based programming. SLASH was recently deployed as the main programming tool in an introductory, undergraduate computer science course for non-majors. Students design objects in the virtual world Second Life and program object behaviors through the code generated by the SLASH environment. By using Second Life as the “laboratory” for the course, students were required to master programmatic constructs through block-generated code and apply the syntax-generated code into Second Life.
An Analysis of Optimizing Impact of Page Design Factor on Cyber Security
Ali Darwish, University of Sharjah
This study investigates the impact of Web page contents design on the effectiveness of browser security indicators. The natural and spontaneous behavior of users’ eyes attention was studied during a security threat to understand what impact the page contents has on noticing security indicators. Sixty participants undertook a usability study that incorporated a controlled phishing attacks and an eye-tracker. The experiment was designed to enable the participants to respond to three phishing and two legitimate Websites. It revealed that there is a strong correlation between current web design and the security cues. This association might be responsible for the increasing number of phishing victims. The study aims to link the results of the experiment to the design recommendations for Web security indicators. These recommendations are based on the elements of design known as composition and the user experience with current designs.
Navigation in a Virtual Environment by Dichotic Listening: Initial Exploration of Audio Cues for BCI Classification
Ashish Dhital, University of Wyoming
Our research is focused on the initial exploration of the possibility of training a Brain Computer Interface (BCI) by using audio cues, instead of motor imagery. We have designed our BCI training and navigation to use audio cues that adhere to the dichotic listening (DL) mechanism so that users have an active choice for interaction or giving commands.
Preparing High School Teachers to Teach Computer Science
David P. Ely, The Ohio State University
Preliminary results of a computer science education study indicate that a successful teacher training project for high school mathematics teachers to teach computer science needs (1) to be in the form of a hands-on workshop with minimal syntax and even less lecture, (2) to use curriculum activities that involve mathematical problem solving. Both programming content knowledge and pedagogical content knowledge of the teachers is being tested. The study is examining a promising Computer Science teacher preparation model and curriculum that is heeding the call of the Transforming High School Computer Science: CS / 10K Project to attain the national goal of training 10,000 well-qualified, K-12 computer science teachers by 2015.
Real-time Visualization of Sentiment Tracking in Twitter
Kathleen Ericson, Colorado State University
Mining Twitter data for sentiment analysis can reveal underlying thought patterns within the Twitter user base. This information can be used to easily poll public sentiment and consumer trends. Previous work has focused on developing algorithms for analyzing tweets, as well as the feasibility of performing this processing in real time. Our work expands on this by moving analysis from file-based Hadoop to stream-based Granules. This should allow us to process streaming tweets more efficiently, allowing us to provide real-time results for visualization. One focus of this work is to allow users to easily visualize analyses on streaming data in real-time, allowing for a deeper understanding of results.
Real-time Visualization A Comprehensive Machine Learning Approach to Predicting and Maximizing Penetration Rates in Earth Pressure Balance (EPB) Tunnel Boring Machines
James Maher, Colorado School of Mines
Tunnel Boring Machines (TBMs) are massive circular drill bits that can be used to excavate small tunnels for utilities or large tunnels for subway systems. TBMs are preferred over traditional blasting and drilling methods because they have the capability to reduce surface deformations and provide a smooth, finished tunnel wall. Predicting the TBM’s penetration rate based on these sensor readings can lead to increased accuracy in project cost estimation. Supervised machine learning models complex systems, based on prior data, to predict future performance. This project implements and evaluates linear regression, polynomial regression, multilinear regression, and support vector regression, to find the best penetration rate prediction model of an EPB TBM.
On-Mote Compressive Sampling in Wireless Seismic Sensor Networks
Marc J. Rubin, Colorado School of Mines
In this extended abstract, we summarize results from simulating compressive sampling as a data reduction technique in a wireless seismic sensor network. We simulated six combinations of sparsity domains and recovery algorithms on real-world passive seismic containing 78 avalanche events. After estimating the original seismic signal, we tested a machine learning workflow to automatically identify avalanche events. The results are promising; with only 30% of the original seismic data, we were able to estimate the original signal well enough to detect nearly 87% of the avalanche events (compared to 90% with full sampling). Lastly, we implemented a novel, lightweight, on-mote compressive sampling algorithm on an Arduino Fio wireless platform and successfully recorded a 220 Hz audio signal.
TreeHouse: Tools for Visualizing and Analyzing Datasets from Large-Scale Phylogenetic Inference
Mark Adamo, Vassar College
Large-scale phylogenetic inference may return sets of thousands of trees, each possibly containing hundreds of taxa. TreeHouse is a phylogenetic tree-querying program that operates on large, highly-compressed sets of trees. I have extended TreeHouse to support multiple modes of visualization and have incorporated support for biological classification data, enabling entirely new hypothesis-testing analyses. New filtering operations allow for trees as well as tree sets to be edited, allowing tree data to be tailored to specialized areas of biological inquiry. Memory optimizations make TreeHouse's memory performance scale much better as the size of trees and trees sets increases.
A Cryptography Module for a Security Video Game
Abdulrahman Almarzooqi, Zayed University
This project is to create a cryptography model for video game that will help teach information security. Similar to a security competition, this game is being designed and implemented to provide students with real world scenarios where they have to make relevant security decisions involving how to apply cryptography. Interviewing Faculty and system administrators as well as downloading and testing related video game softwares provided great insight for the delopment of these prototypes. Tests with students and teachers will provide valuable insight to move from a prototype to a final product.
Teaching Network Security with a Video Game Simulating Security Competitions
Fatma Y. Beshwari, Zayed University
This project is to create a network module of a video game that will help teach information security. This game being designed is similar to a security competition, where students need to take quick actions to defend their resources. The target audience is university students preparing for a security competition or taking a class on penetration testing. Faculty and system administrators were interviewed to design the scenarios. The first prototype is being implemented in Game maker and after classroom testing we will implement it in Unity.
Metagenomic Data Analysis using Clustering
Sulochana Bramhacharya, Hiram College
Recombinant DNA technology and the decrease in sequencing costs have revolutionized the area of genetics. One of the biggest challenges right now is the analysis of gene sequences, gene clusters, and differences in composition of these clusters. Our program is a simple tool designed to cluster the gene sequences from samples and identify the outliers in two similar clusters from different samples. We took 16S gene sequences from two similar samples of different environmental conditions and clustered them using a greedy incremental approach based on percent identity. We compared each cluster obtained from the first sample with each cluster in the second sample and determined two corresponding cluster pairs from both samples. If any of the clusters correspond, two different sets of data are obtained, i.e. one set of similar sequences in both clusters and another set of dissimilar sequences. The set of dissimilar sequences, also known as outliers, is useful to study about the factors causing the change in cluster composition in response to environmental changes. This program is aimed to be used by the students in their classroom and researchers.
An Improved User Interface for the Corona Project
Coty Collins, Saint Bonaventure University
Corona is a system that automates the process of creating JUnit tests for beginning programmers. In previous versions, users needed to know an arcane syntax to create a test. To address this, an improved user interface was developed. Syntactic overhead is reduced via a tree-like structure that allows users to specify objects. Parameters in a collection are also represented in the tree as children for the nodes in their containing class. Along with other improvements, the new user interface for Corona frees users of syntactic concerns and allows them to focus on input/output pairs; early studies suggest users are likely to generate more tests.
Debugging Using a Verifier
Carrie Eisengrein, Clemson University
Today, people normally find errors in programs by running tests. This can cause problems because there are occasions when tests cannot fully check every part of the system. Testing is not foolproof. The main goal in my area of study is to explore how to diagnose and fix errors in a program using verification conditions generated by the system, RESOLVE.
Automation of the Pre-Registration Process at Baldwin Wallace University
Aric F. Gady, Baldwin Wallace University
In an effort to aid student retention, Baldwin Wallace University implemented a pre-registration service for incoming first-year students that provides them with a pre-determined schedule as a starting point for planning their first semester. This program requires significant man-hours to create more than 600 schedules each year. The aim of this project is to automate the summer chairs’ registration process for those incoming first-year students using simulated annealing. We have found that not only can the schedules be completed in less time, but also that the schedules may be of higher quality by some measures.
Porting Xinu to Raspberry Pi
Farzeen Harunani, Marquette University
As the field of computer science expands and becomes increasingly more complicated, two things happen. First, hardware and software get outdated regularly, and staying "up to date" can be expensive. Second, in terms of teaching computer science, it can be difficult for newcomers to the field (mainly students) to grasp the underlying concepts due to modern computing's ever-increasing system complexity. By merging an inexpensive piece of hardware with a simple Unix-like operating system, we can create a teaching tool that is accessible to anyone in terms of usability as well as monetary cost. This project proposes to utilize the XINU operating system in conjunction with the Raspberry Pi development board to achieve this goal.
The Robustness of Medical Decisions with Noisy Data - An analysis of the robustness of eGFR calculations using unreliable inputs
Nathan Lapinski, University of Colorado
Errors in medical data are one of the leading causes of fatalities in hospitals worldwide, leading to thousands of preventable deaths each year. In an effort to reduce this problem, many healthcare institutions have implemented clinical decision support software to aid practitioners in the patient care process. However, studies have shown that such software systems can actually lead to an increase in data errors, due in part to a lack of internal data validation mechanisms. We studied the role of a clinical decision support system in the diagnosis of renal disease and established that up to 27% of decisions made using such measures may be inaccurate. We propose a software solution to this problem that assesses the robustness of patient data inputs before they are used in clinical decision support software.
Digital Systems Test Bench
Amir Sabet, AASTMT
The purpose of this system is to test the digital circuits implemented by the students. The user either the student or the instructor supplies the design of the digital circuit in the form of HDL like text, for the simulation purpose. The main inputs and outputs of the circuit are connected to the computer for testing and error prediction. The proposed system software runs all the test inputs and compares the output from the circuit to the output from the simulation. Using the mismatching cases a prediction of the faulty connections is provided to assist the debugging of the circuit.
Enabling a Resource Limited Robot to Formulate Complex Plans
Demetrius Taylor, Lamar University
This study investigates the use of an existing planning system in conjunction with the INTERRAP architecture for autonomous agents to formulate plans for a resource limited task-oriented autonomous robot. The study creates a system that interfaces Dana Nau's hierarchical task network (HTN) planner with the local planning layer (LPL) of Jorg Muller’s INTERRAP architecture. Since SHOP2 cannot run on the resource limited robot, a PC running SHOP2 formulates plans, and those plans are communicated to the LPL and executed by the robot. This enables the robot to achieve goals that require complex plans and permits reuse of a high-quality award-winning planner. The study presents the design, implementation highlights and a demonstration of an application of the INTERRAP-SHOP2 system.
Dimitri Wijesinghe, Vassar College
MCMCTree is a popular program for calculating the branch lengths of phylogenetic trees. It operates by taking an evolutionary tree and uses fossil data to estimate the origin times of each species, via the Markov Chain Monte Carlo probabilistic model. These calculations are computationally intensive and large datasets invite system failure. We have implemented a system of checkpointing for MCMCTree. With this system, MCMCTree can begin a calculation and save data as it goes along. If the system fails, calculations can be restored from the last checkpoint instead of starting from scratch. This infinitely increases the range of data the system can process. With checkpointing, MCMCTree can process trees of greater scope, and view evolution on a much broader scale than before.
Wireless Sensor Network Solution for Monitoring Earth Dams
Vladimir Yaremenko, Colorado School of Mines
For this project, I have written the firmware for a custom wireless device (mote) developed by our research group to monitor earth dams. A network of these motes is intended to replace current monitoring technology which is very limited in its data acquisition rate. With these motes placed in key points around the dam they will be able to provide a constant stream of data to a central location. Geophysicists will then be able to monitor the data and quickly detect any abnormalities. After implementing some form of time synchronization on these motes, they will be ready for full scale tests versus the current measuring equipment used for earth dams.