2021 Prospective Student Visit Day and Graduate Research Symposium [Hybrid]

2021 Prospective Student Visit Day and Graduate Research Symposium [Hybrid] promotional image

The Computer Science (CS) department welcomes prospective students interested in our graduate programs to our annual Prospective Student Visit Day on Fri, Nov 5. We are looking for strong students with diverse backgrounds to join our MCS and PhD programs. A wide variety of research areas is represented by our world-class faculty including algorithms, computational epidemiology, distributed computing, human-computer interaction, machine learning, massive data algorithms and technology, mobile computing, networks, programming languages, text mining, security, and virtual reality. If you intend to join us, please RSVP our graduate coordinator Ms. Sheryl Semler at sheryl-semler@uiowa.edu, by October 31.

An important part of the prospective Student Visit Day is the 7th Iowa Computer Science Graduate Research Symposium (2021), in which senior PhD students will present talks showcasing their current research. This will be followed by a "keynote" by a CS faculty member. Talks are intended for a wide audience with interest in CS, including CS juniors and seniors. The talks presented by current CS graduate students at the Symposium are excellent examples of the exciting CS research taking place here on the UI campus!


All times CT [Conversion to your timezone possible here if needed]

Friday, Nov 5, 2021

Morning Session (By invitation)



Overview of Graduate Programs: Led by Professor Kasturi Varadarajan, Director of Graduate Studies

Faculty Research Presentations by Professors Juan Pablo Hourcade, Garrett Morris, Tianbao Yang, and Octav Chipara


Graduate Student Research Symposium Sessions [Online meeting URL; Passcode: 583807]


Student Sessions


Jon Rusert, PhD Student

"Don’t sweat the small stuff": Sampling Text to Shield against Adversarial Attacks


Dhruv Vyas, PhD Student

Hearing Aid Personalization Systems from Computer Science Perspective


M Hammad Mazhar, PhD Student

On the "Safety Problem" of the Internet of Things (IoT) Environments


Jeehan Malik, PhD Student

Understanding Age-Related Differences in Technology Use for Physical Mobility






Speaker: Professor Supreeth Shastri

Legalizing Personal-data Systems



Jon Rusert

Title: “Don’t sweat the small stuff”: Sampling Text to Shield against Adversarial Attacks

Jon Rusert portrait

As the internet grows, more websites and companies have been leveraging deep learning in the form of text classification systems to help quickly identify and categorize text. This is useful for identifying sentiment towards a product or person, flagging misinformation and hate speech, and much more. In recent years, however, researchers have shown these systems are vulnerable to adversarial attacks. That is, other deep learning systems learn to change the text in a way which causes the text classifier to misclassify the text, while the original meaning of the text is still kept intact. These adversarial attacks (or obfuscators) have found great success in greatly dropping classification accuracy while retaining semantic integrity across a variety of text classification datasets. To maintain original meaning, current state-of-the-art (SOTA) obfuscators change as little text as needed to cause a misclassification. We take advantage of this by introducing Sample Shielding. Since only a small portion of text is changed, we take samples of that text which have a chance of not including the changed text. We find that classifying the samples results in a large decrease in the effectiveness of the adversarial attacks. We examine our methodology against two SOTA obfuscators across four classification datasets. We find that our shielding can reduce an attack’s effectiveness from 99% effective to 15% effective. That is, a classification system which achieved 91% accuracy and dropped to 1% accuracy against adversarial texts, can still achieve 78% via sample shielding against the same adversarial text. Furthermore, our shielding method is obfuscator agnostic and it doesn’t need to know the inner workings of the obfuscator to defend against the adversarial text it produces. Finally, we build on top of our shielding method and point to future directions in which this technique could be expanded.

5th Year PhD student | Advisor: Padmini Srinivasan | Areas of research: Web & Text Mining, Natural Language Processing, Privacy, Censorship, Obfuscation, Adversarial Debiasing

Dhruv Vyas

Title: Hearing Aid Personalization Systems from Computer Science Perspective

Dhruv Vyas - from LinkedIn

Over the past two decades, new and increasingly complex signal processing algorithms have been developed for hearing aids. A key challenge to harnessing the full potential of these algorithms is to tune the numerous parameters that control their behavior. A critical challenge is to provide users with an easy-to-use method for personalizing many parameters which control sound amplification based on their preferences. In this talk, we present a novel approach to personalize hearing aids that provides a higher degree of personalization than existing methods by using user feedback more efficiently. We perform a 35-user study to understand the factors that affect user preferences and evaluate the efficacy of state-of-the-art personalization systems. We show that the proposed approach can find the best configuration with a median of 25 comparisons, reducing by half the comparisons required by the best baseline.

6th year PhD student | Advisor: Octav Chipara | Areas of research: Healthcare systems, Audiology, Reinforcement learning, Data analytics

M Hammad Mazhar

Title: On the "Safety Problem" of the Internet of Things (IoT) Environments


M. Hammad Mazhar portrait

The rapid adoption of the Internet of Things (IoT) ecosystem in various contexts (residential, industrial, and commercial) makes these IoT systems a lucrative target for malicious manipulation. To make matters worse, IoT devices often co-reside in the same physical space as the user, giving these devices access to sensitive information about users and as well as the opportunity to negatively impact users physical safety. In addition, as IoT devices become more sophisticated, the ecosystem becomes more complex in terms of how devices interact with each other leading to a rise in security and privacy issues in deployments. In this talk, we show that defending against most of these security and privacy concerns boils down to the ability of solving "the safety problem'' in IoT environments. We start by precisely defining the "safety problem in IoT environments", which is essentially capturing the requirement of regulating IoT environments so that they behave according to user expectations. We then look at how prior work attempted to solve this problem and outline their limitations. We then present our work MITOSYS, which addresses these limitations by providing a platform-agnostic solution while solving the safety problem. Finally, we outline future work in this space along with the challenges expected in such work.

6th year PhD student | Advisor: Omar Chowdhury | Areas of research: IoT security and privacy, policy enforcement

Jeehan Malik

Title: Understanding Age-Related Differences in Technology Use for Physical Mobility

Jeehan Malik

Understanding age-related differences in technology use is important because older adults are often underrepresented in technological innovations. This report provides an overview of research on assistive technology for physical mobility for older adults, and research on emerging technology to highlight the gaps where age-related differences have not been explored. It further examines the specific context of street-crossing, a common physical mobility task. Governmental agencies and vehicle manufacturers are exploring sophisticated technology for pedestrian safety. Vehicle-to-Pedestrian (V2P) technology is one such technology, and it is important to understand how pedestrians respond to different types V2P communication. This presentation gives an overview of how V2P communications, using external Human-Machine Interfaces (eHMI), smartphones, and Augmented Reality (AR), are received by older and younger adults.

4th year PhD student | Advisor: Kyle Rector | Areas of research: HCI, digital accessibility, aging

Keynote Speaker Supreeth Shastri

Title: Legalizing Personal-data Systems


In recent years, many societies are granting a new fundamental right to the people—the privacy and protection of personal data. While regulations such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) are already in effect, modern computing systems are struggling to comply with them. In this talk, I will show how these laws invalidate computing principles and practices that have decades of precedence, and why it is challenging to legalize personal-data systems. The focus of my research has been to bridge this gap between (legal) intention and (computing) implementation. I will share the work that my collaborators and I have been doing in this space, and identify some of the interesting open problems.

Supreeth Shastri portrait

Bio: Supreeth Shastri is an Assistant Professor of computer science at the University of Iowa. His research interests lie in experimental computer science with an emphasis on systems and networking. He received his PhD from the University of Massachusetts at Amherst and MS from Columbia University. Previously, he held a Postdoctoral Fellowship at the University of Texas at Austin, and spent several years writing system software at ARM and Cisco Systems. His research has been adapted into best practices in the industry, deployed at the U.S. Federal Aviation Administration, integrated into courses at a dozen universities across the US and Europe.

Friday, November 5, 2021 1:30pm to 5:00pm
University Capitol Centre
200 South Capitol Street, Iowa City, IA 52240
View on Event Calendar
Individuals with disabilities are encouraged to attend all University of Iowa–sponsored events. If you are a person with a disability who requires a reasonable accommodation in order to participate in this program, please contact Sheryl Semler in advance at 319-335-0707 or sheryl-semler@uiowa.edu.