Undergraduate Research Week 2021
April 19-23 is both Undergraduate Research Week and the Iowa Center for Research by Undergraduates' 2nd ESURF (Spring Undergraduate Research Festival).
This year, SURF is going to run a little bit differently than it has ever before. ICRU hopes to provide a valuable presentation experience for all of its undergraduates. We have designed this event with the following thoughts:
- Presenting is an incredibly valuable experience for developing researchers;
- Zoom fatigue is real and present;
- Schedules for presenters and audience members are hectic;
- Virtual presentations are now a common place part of life.
- Undergraduate Research Week is April 19-23, 2021.
Given these few thoughts, ESURF will occur during Undergraduate Research Week and will consist only of 3-5 minute, pre-recorded presentations. No live events will occur. (Students who wish to present live should consider the ICRU Lecture Series.) Videos will be released throughout the week on ICRU's website and social media channels. Feedback from commentators (at the undergraduate, graduate, post-doc, and faculty level) will be provided if students wish to receive feedback.
ESURF 2021 will feature Academic Year 2020-21 research in our department including:
Improving the Accessibility of Dynamic Trees for Blind CS Students
Majors: Computer Science, English
Mentor(s): Brandon Myers, Kyle Rector (Computer Science)
Trees are used throughout computer science to visualize algorithms and data structures. Unfortunately, trees are largely inaccessible to blind CS students due to their graphical formats, which are incompatible with screen readers.* Although there has been prior work done to make static, or unchanging, trees accessible to blind CS students, there remains work to be done in making dynamic, or changing, trees accessible.
There is a possibility that presenting trees in alternative formats, i.e. textually and auditorily instead of visually, can help blind CS students better understand the changes that occur in trees.
To determine how alternative formats might help, we plan to conduct interviews with blind CS students (past or present) to investigate:
- how they currently learn about trees,
- what types of tools or resources they are using, and
- what new tools or resources they would be open to using.
* Screen readers are tools that blind and visually impaired individuals rely on to read and interpret the data on their computer screens.
Making Aerobic Exercise More Accessible to People with visual impairments
Majors: Mathematics, Computer Science
Mentor(s): Kyle Rector (Computer Science)
Currently there are many barriers that prevent people with visual impairments from accessing aerobic exercise. First, exercises that have been adapted for people with visual impairments are niche or regionally popular. In addition, there is stigma in group exercise environments in the form of excessive attention placed on disability as well as patronizing behavior which lead people with visual impairments to be dissatisfied. Finally existing technologies that provide aerobic exercise feedback lack nuance and require a visual reference. We propose a system designed for step aerobics exercise that can give a participant with visual impairments nuanced audio feedback that can be used discretely to avoid attracting attention to disability.
Machine Learning in the Smart Home
Major: Computer Science
Mentor(s): Brandon Myers (Computer Science)
Internet of Things (IoT) devices have surged in popularity over the last several years. Much of this can be credited to common devices in the home becoming smart and internet-connected. These devices include everything from lightbulbs to door locks, introducing new functionality and features. While these features often provide higher convenience to users, such as using your mobile device to control your home equipment from anywhere, it can also be cumbersome to forgo traditional control, like flipping a lightswitch. Voice Assistants like Google Assistant or Amazon Alexa have been at the forefront of controlling smart home devices with speakers and microphones in the home. While these solutions do allow a user to control their home without their mobile device, they can be time consuming or annoying when not working properly.
We propose using machine learning (ML) techniques to allow a home to utilize past behavior and make adjustments to smart devices without the user having to do anything. Our approach leverages time and presence-based factors to create a system that can operate an entire home independently of human intervention. We begin by collecting data regularly which includes the time, day, and presence of users among other features. We then select 14 types of ML models and make eliminations until just one remains. Then, we optimize the final model and evaluate it using real-time data. Lastly, we deploy the model to independently control the home.
The University of Iowa Libraries has opened the undergraduate library research award to all students, whether they are participating in ICRU's Undergraduate Research Festivals or not. This is to encourage more students to apply for this award. The deadline is April 30th.