News
School of Computing Celebrates Student Scholars Presenting Today
April 13, 2022
Student Scholars Day is today, April 13th. The event is being held in Kirkhof Center as well as the Henry Hall Atrium from 9:00 a.m. until 5:00 p.m. Please consider joining to see several computing students present as well as students from other undergraduate majors. This is a day to celebrate the hard work our student researchers have been doing over the course of the semester and even longer.
The computing students’ abstracts are listed below along with the information on where to find them stationed at the event.
Finding Malicious Malware Using Twitter Scanner
Morgan Hamlin
Mentored by Erik Fredericks
Kirkhof Center, GRR 001
10:00-11:00
Zoom has become an increasingly popular application. This has led to an increase in cyber criminal activity. Cybercriminals have been utilizing social engineering tactics to persuade users to download compromising software. Their agenda is to gain access to sensitive information and exploit your system. I will be utilizing a Python bot to find malicious malware posted on Twitter under the domain name of Zoom. This bot can scan and identify potential threats, with the idea being to bring awareness to phishing scams and help mitigate dangerous Zoom downloads on Twitter.
A New Space-Filling Curve
Elise Dettling
Mentored by Christian Trefftz
Kirkhof Center, GRR 055
11:00-12:00
Space-filling curves have many applications in scientific computing. Most space-filling curves are displayed using cartesian grids and they work in conjunction with traditional quadtrees. We have proposed a new kind of curve, the LTD curve. We have written a grammar that can generate different levels of recursion of the curve. We have chosen to use barycentric coordinates instead of traditional cartesian coordinates, and we have devised a new flavor of quadtree designed for triangular grids. Previous work on this subject has been surveyed by Michael Bader in his book Space-Filling Curves: An Introduction with Applications in Scientific Computing.
Identifying Tradeoffs in Benefits of Interaction and Visualization Techniques for Analytic Tasks
Calob Horton and Shweta Dnyaneshwar Terkar
Mentored by Michelle Dowling
Kirkhof Center, GRR 005
1:00-2:00
In visual analytics research, new visualization or interaction techniques are typically developed to support specific analytic tasks, such as clustering observations of high-dimensional data. User studies on these new techniques seek to demonstrate their effectiveness on specific tasks. Developers of visual analytics systems then seek to combine these techniques to support more tasks, such as characterizing distributions and clustering observations. In this example, the effectiveness of a clustering technique on characterizing distribution tasks was not evaluated and may be detrimental to this task. Therefore, a different technique for this task may be preferred. As such, we promote a broader perspective when evaluating the effectiveness of a new technique so that the tradeoffs in the benefits of using such techniques across a wide range of analytic tasks can be understood. This presentation demonstrates how to complete such an evaluation on the techniques employed by SIRIUS.
Mitigating the Impact of Run-Time Software Engineering
Abigail Diller
Mentored by Erik Fredericks
Henry Hall, Atrium 025
1:00-2:00
Cyber-physical systems (CPSs) experience uncertainty in all phases of the software development life cycle, generally stemming from real-world concerns such as minimizing the impact of environmental interactions and ensuring the safety of human participants. These systems are often described as safety-critical and therefore must continuously exhibit correct behaviors. Run-time testing can provide assurance that system requirements are satisfied even in the face of uncertainty. Previously we introduced GreenRoom, a proof-of-concept experimental CPS testbed for monitoring the impact of run-time testing on power consumption of a target CPS. This project extends GreenRoom’s capabilities to enable deeper studies into monitoring power consumption by including additional mechanical components. We also intend to investigate how search-based testing techniques can extend CPS abilities, and to monitor their impact on power consumption at run time.
Detecting Face Mask Usage Trends in Social Media with Machine Learning
Seth Ockerman
Mentored by Erin Carrier
Kirkhof Center, GRR 002
2:00-3:00
The use of face masks to prevent disease spread among the general population has become widespread due to the COVID-19 pandemic. The ability to accurately detect and monitor the trends in face mask usage is crucial to understanding and predicting hotspot areas for both current and future pandemics. This work investigates the detection of face masks in social media images using deep learning. This project creates a social-media-based face mask image dataset that reflects the scale needed for deep learning and the diversity (mask types, positions of people, and ethnicity) of real life. We have gathered approximately 120k images containing people from different cities. Mechanical Turk is used to label the images based on the presence of a face mask. Using this dataset, we train a CNN model to detect the presence of face masks in social media images and compare the results to existing datasets. We then deploy our model to detect trends in face mask usage in Los Angeles over time.
Exploring Fast Emulators for Long-Running Mars Simulation Code
Marc Tunnell
Mentored by Erin Carrier and Nathan Bowman
Kirkhof Center, GRR 013
4:00-5:00
The NASA Ames Global Climate Model (GCM) software has been in steady use at NASA for decades and was recently released to the public. This model simulates the complex interactions of different weather cycles that exist on Mars, namely the Dust Cycle, the CO2 Cycle, and the Water cycle. It is used by NASA to help understand their empirically observed data through the use of sensitivity studes. However, these sensitivity studies are computationally taxing, requiring weeks to run. To address this issue, we have developed a surrogate model using Gaussian processes that can emulate the output of this model with relatively small amounts of data in a reduced amount of time (on the order of minutes). We demonstrate the effectiveness of our emulator using backward error analysis.