- Gradstudents - mailman.cs.vt.edu

Mounika Abbineni's MS defense, 1:30 pm 5/6
by Hooshangi, Sara 09 May '25

09 May '25

Hello, I am delighted to invite our CS community to Mounika Abbineni's MS defense tomorrow (Tuesday). Please see details below: Title: Mixed Methods Study of How Computer Science Students Make Elective Course Choices Date: Tuesday May 6, 2025 Time & Location: 1:30 pm, on Zoom Zoom: https://virginiatech.zoom.us/j/88202275099?pwd=3CbKfvTnvJU3J3bIoMKpQQUWuC42… Committee Members: Sara Hooshangi (Chair), Jin-Hee Cho, Brandi Weiss (GWU) Abstract: Elective course selection is an important part of the undergraduate Computer Science learning experience. These courses often shape a student’s future career along with teaching them new skills and forming their niche in the vast world of Computer Science. While computer science students in Virginia Tech have a designated checklist along with required courses that they are supposed to take, the curriculum offers them a set of technical electives where they are able to learn topics of interest in the vast field of Computer Science. By the time of graduation most Computer Science students have taken about three or four elective courses, where they are able to explore their interests and strengths in Computer Science. This mixed-methods study utilized responses from undergraduate, graduate, and recent alumni from Computer Science, Computer Engineering, and Data Analytics majors at Tech. For the qualitative analysis, a thematic analysis was done from 14 student interviews. A quantitative analysis was performed through survey responses. A statistical analysis of survey responses was done in order to analyze the data and make conclusions on the factors that influence student choices of CS electives. It was observed that as students gain more exposure, they approach elective decisions with a focus on the impact of their career and the utility it offers. Students consistently prioritized taking courses that built their skill set and added value to their career goals. The only gender difference found in how students choose electives was in how important course format factors were, like class timing, modality, and grade distribution. Secondly, students in different years of study prioritized the value of professor-related and course format factors differently in their elective choices. This research aims to help young students find their grounding coming into college so they are aware of how to make better course choices. It also aims to introduce better academic advising strategies and curriculum development such that a students' experiences are better in college with easier transition into a successful and enjoyable work life.} ---------------------------------- Sara Hooshangi, Ph.D. Collegiate Associate Professor Director of Master of Engineering Program Department of Computer Science Virginia Tech

1 2

Reminder: Seminar TODAY: May 9th, 11am: Title: "Large Language Models: Opportunities, Challenges and Pitfalls"
by Kantarcioglu, Murat 09 May '25

09 May '25

Dear All, Please see the details below for an upcoming seminar scheduled for May 9th at 11:00 AM in Room 1100 , Torgersen Hall (Corrected Location). We encourage you to attend if your schedule allows. ------------------------------------------------------------------------------------------------------------------------------------- Title: "Large Language Models: Opportunities, Challenges and Pitfalls" Abstract: Ever since ChatGPT was released, Large language models (LLMs) have captured the imagination of researchers and users worldwide and led to a capacity race. We examine their utility for detecting and explaining deceptive attacks such as fake news, fake product reviews, and phishing. Then, we consider the challenges in the context of the detectability of LLM-generated content. We conclude with our recent outside-the-class experiences with LLMs at the University of Houston and at the Naval Surface Warfare Center (NSWC), Dahlgren. Directions for future research and insights into LLMs revealed by our research will be discussed including the consequences of aligning LLM models for different purposes, such as 1) ability to mimic humans on natural language tasks and 2) safety. (This talk is based on joint work with Dainis Boumber, Nachum Dershowitz, David Marchette, and Bryan Tuck) Bio: Rakesh Verma is a Professor of Computer Science at the University of Houston (UH), where he pioneered a Security Analytics course over a decade ago. He received UH's 2013 Lifetime Faculty Mentoring Award for his extensive mentoring of undergraduate students in research. He is the author of Cybersecurity Analytics (CRC Press, 2019) with David Marchette. He is a cybersecurity section associate editor of the Frontiers in Big Data journal. He has been organizing the ACM Annual International Workshop on Security and Privacy Analytics (IWSPA) since 2015. He is currently the PC Co-chair of ICISS 2025. His research interests include formal methods, algorithm design and analysis, and computer security. He is on sabbatical at NSWC Dahlgren for the 2024-2025 academic year. See Cybersecurity Analytics (CRC Press, 2019) at https://www2.cs.uh.edu/~rmverma/book.html Date and Time: May 9th at 11:00 AM in Location: Room 1100 Torgersen Hall (Corrected Location). Zoom info: https://virginiatech.zoom.us/j/86821347393?pwd=IH4ntEcThTIaojr0xYcxbxbyc828… Meeting ID: 868 2134 7393 Passcode: 318285 ------------------------------------------------------------------------------------------------------------------------------------- Best Regards, Murat ==================================== Murat Kantarcioglu, Ph.D. Professor, and CCI Faculty Fellow Virginia Tech, Department of Computer Science www.kantarcioglu.net ====================================

1 0

Sophia Stil's MS Thesis Defense - 5/13 7pm ET
by Rho, Eugenia 08 May '25

08 May '25

Hi All, You are cordially invited to the Master's Thesis Defense of Sophia Stil. The details are as follows: Date: 05/13/2025 Time: 7:00 PM ET Title: Understanding the Applications of AI for Autistic STEM Student Mentorship Location: 3218 Gilbert Place Zoom: https://virginiatech.zoom.us/j/81187632432 Committee: Eugenia Rho (Chair), Myounghoon Jeon, Sang Won Lee Abstract: Autistic college students in STEM face significant challenges in employment, such as low-pay and hiring discrimination compared to their neurotypical peers, despite existing mentorship programs for career development. This study investigates how artificial intelligence (AI) can be leveraged to enhance STEM mentorship for autistic students. Specifically, it examines students’ acceptance of AI as a complementary support tool alongside human mentors, their previous mentorship experiences, their experiences and perceptions of AI, as well as their deeper visions for future AI integration in mentorship. Chapter 2 presents findings from an exploratory online survey of 203 autistic students, offering insights into their perceptions of AI-augmented mentorship, including perceived benefits, limitations, and areas of concern. Building on these results, Chapter 3 describes a follow-up interview study with 15 participants that explores their current academic and professional experiences typically supported by mentorship and their vision for how AI could support key mentorship activities. Thematic analysis reveals a strong desire for integrated, intelligent tools to facilitate online communication, time management, and emotional well-being, alongside valid concerns regarding the limitations and potential harms of current AI systems. The findings inform design considerations for future AI-supported mentorship and underscore the importance of accessibility and inclusion in the development of such systems. We conclude with recommendations for the design of future mentorship programs and AI to support the unique needs of autistic students in STEM. Eugenia Rho<https://eugeniarho.com/> Assistant Professor<https://eugeniarho.com/> Computer Science | Virginia Tech<https://cs.vt.edu/> [cid:a4758e9e-9d99-4f67-a691-0edc4c4e3b8e]<https://eugeniarho.com/sail/>

1 0

Reminder: Seminar: May 9th, 11am: Title: "Large Language Models: Opportunities, Challenges and Pitfalls"
by Kantarcioglu, Murat 08 May '25

08 May '25

Dear All, Please see the details below for an upcoming seminar scheduled for May 9th at 11:00 AM in Room 1100 , Torgersen Hall (Corrected Location). We encourage you to attend if your schedule allows. ------------------------------------------------------------------------------------------------------------------------------------- Title: "Large Language Models: Opportunities, Challenges and Pitfalls" Abstract: Ever since ChatGPT was released, Large language models (LLMs) have captured the imagination of researchers and users worldwide and led to a capacity race. We examine their utility for detecting and explaining deceptive attacks such as fake news, fake product reviews, and phishing. Then, we consider the challenges in the context of the detectability of LLM-generated content. We conclude with our recent outside-the-class experiences with LLMs at the University of Houston and at the Naval Surface Warfare Center (NSWC), Dahlgren. Directions for future research and insights into LLMs revealed by our research will be discussed including the consequences of aligning LLM models for different purposes, such as 1) ability to mimic humans on natural language tasks and 2) safety. (This talk is based on joint work with Dainis Boumber, Nachum Dershowitz, David Marchette, and Bryan Tuck) Bio: Rakesh Verma is a Professor of Computer Science at the University of Houston (UH), where he pioneered a Security Analytics course over a decade ago. He received UH's 2013 Lifetime Faculty Mentoring Award for his extensive mentoring of undergraduate students in research. He is the author of Cybersecurity Analytics (CRC Press, 2019) with David Marchette. He is a cybersecurity section associate editor of the Frontiers in Big Data journal. He has been organizing the ACM Annual International Workshop on Security and Privacy Analytics (IWSPA) since 2015. He is currently the PC Co-chair of ICISS 2025. His research interests include formal methods, algorithm design and analysis, and computer security. He is on sabbatical at NSWC Dahlgren for the 2024-2025 academic year. See Cybersecurity Analytics (CRC Press, 2019) at https://www2.cs.uh.edu/~rmverma/book.html Date and Time: May 9th at 11:00 AM in Location: Room 1100 Torgersen Hall (Corrected Location). Zoom info: https://virginiatech.zoom.us/j/86821347393?pwd=IH4ntEcThTIaojr0xYcxbxbyc828… Meeting ID: 868 2134 7393 Passcode: 318285 ------------------------------------------------------------------------------------------------------------------------------------- Best Regards, Murat ==================================== Murat Kantarcioglu, Ph.D. Professor, and CCI Faculty Fellow Virginia Tech, Department of Computer Science www.kantarcioglu.net ====================================

1 0

PhD dissertation final defense
by Yao, Danfeng 05 May '25

05 May '25

Hello everyone, I am pleased to invite you to the Dissertation Defense of Alexander Kedrowitsch. The details are as follows: Date: 05/06/2025 Time: 11:00 AM Title: New Security Paradigms for Spacecraft and Networks: Metrics, Testbeds, and Scalable Solutions Place: 4001 Gilbert Place Zoom: Meeting ID: 842 260 4413, Passcode: 4413 Committee: Daphne Yao (Chair), Jonathon Black (Co-chair), Jin-Hee Cho, Wenjie Xiong, Samuel Jero Abstract: The rapid expansion of commercial spaceflight, driven by reduced launch costs and increased private investment, has transformed the landscape of space-based communications. Proposed mega-constellations comprising thousands of satellites promise global connectivity, linking directly to handheld devices on Earth. However, as government and commercial actors push further into space, the emerging heterogeneous networks that interconnect these systems face significant challenges in security and resilience—areas often discussed but infrequently demonstrated. This dissertation addresses these challenges through four primary contributions. First, it presents a novel logical topology for mega-constellations that enables scalable and resilient dynamic routing, significantly reducing network and computational overhead in large-scale satellite networks. Then, it introduces techniques for enabling robust, secure communication on energy-harvesting spacecraft, balancing security requirements with constrained power budgets. Additionally, this work evaluates the suitability of NIST-certified encryption algorithms for deep-space platforms, ensuring compliance with established standards while considering space-specific constraints. The dissertation also presents several holistic frameworks for systematic performance evaluations of dynamic network routing resilience and intermittent cryptography in space environments. Lastly, the dissertation describes contributions made to several open-source network emulators, enhancing their utility for future space networking research. Together, the contributions of a scalable logical topology, energy-aware NIST-compliant cryptographic mechanisms, holistic evaluation frameworks, and enhanced open-source emulators enable practical, standards-aligned research and deployment of next-generation satellite communications.

1 0

Master's Thesis Defense -- Jan Michalak
by David-John, Brendan 05 May '25

05 May '25

Hello everyone, You are cordially invited to the Master's Thesis Defense of Jan Michalak. The details are as follows: Date: 05/13/2025 Time: 3:00 PM Title: Biometric Leakage from Generative Models and Adversarial Iris Swapping for Spoofing Eye-based Authentication Place: 3211 Gilbert Place Committee: Brendan David-John (Chair), Bimal Viswanath, Bo Ji Abstract: This thesis investigates the vulnerability of generative models trained on biometric data and explores digital spoofing attacks on iris-based authentication systems representative of AR/VR environments. We first explore how diffusion models trained on biometric data can memorize and leak iris images. Next, we evaluate the effectiveness of Cross-Attention GANs for iris-swapping attacks, demonstrating their ability to enable presentation attacks that spoof iris-recognition systems. Our experiments across several standard iris and VR datasets have an attack success rate of 100% within similar domains and generalize across domains with rates as high as 70%. Our findings highlight the need to consider vulnerabilities in biometric systems and strengthen defenses against digital presentation attacks produced by generative models. Cheers, Dr. Brendan David-John Assistant Professor, Computer Science Virginia Tech private-eye.app<http://private-eye.app/>

1 0

Pizza Delay for Picnic
by Moussad, Bernard 02 May '25

02 May '25

Hey guys, There was a mix-up on our pizza order so it'll be at least 30 minutes until we can pick it up. Until then, we'll have some drinks and cookies for folks until it's ready. Sincerely, Bernard Moussad PhD student CS@VT Phone: 256-668-7571 Website: Bernym12.github.io<https://Bernym12.github.io>

1 0

Seminar: May 9th, 11am: Title: "Large Language Models: Opportunities, Challenges and Pitfalls"
by Kantarcioglu, Murat 02 May '25

02 May '25

Dear All, Please see the details below for an upcoming seminar scheduled for May 9th at 11:00 AM in Room 100, Torgersen Hall. We encourage you to attend if your schedule allows. ------------------------------------------------------------------------------------------------------------------------------------- Title: "Large Language Models: Opportunities, Challenges and Pitfalls" Abstract: Ever since ChatGPT was released, Large language models (LLMs) have captured the imagination of researchers and users worldwide and led to a capacity race. We examine their utility for detecting and explaining deceptive attacks such as fake news, fake product reviews, and phishing. Then, we consider the challenges in the context of the detectability of LLM-generated content. We conclude with our recent outside-the-class experiences with LLMs at the University of Houston and at the Naval Surface Warfare Center (NSWC), Dahlgren. Directions for future research and insights into LLMs revealed by our research will be discussed including the consequences of aligning LLM models for different purposes, such as 1) ability to mimic humans on natural language tasks and 2) safety. (This talk is based on joint work with Dainis Boumber, Nachum Dershowitz, David Marchette, and Bryan Tuck) Bio: Rakesh Verma is a Professor of Computer Science at the University of Houston (UH), where he pioneered a Security Analytics course over a decade ago. He received UH's 2013 Lifetime Faculty Mentoring Award for his extensive mentoring of undergraduate students in research. He is the author of Cybersecurity Analytics (CRC Press, 2019) with David Marchette. He is a cybersecurity section associate editor of the Frontiers in Big Data journal. He has been organizing the ACM Annual International Workshop on Security and Privacy Analytics (IWSPA) since 2015. He is currently the PC Co-chair of ICISS 2025. His research interests include formal methods, algorithm design and analysis, and computer security. He is on sabbatical at NSWC Dahlgren for the 2024-2025 academic year. See Cybersecurity Analytics (CRC Press, 2019) at https://www2.cs.uh.edu/~rmverma/book.html ------------------------------------------------------------------------------------------------------------------------------------- Best Regards, Murat ==================================== Murat Kantarcioglu, Ph.D. Professor, and CCI Faculty Fellow Virginia Tech, Department of Computer Science www.kantarcioglu.net ====================================

1 0

CS Graduate Program Newsletter: May Edition
by Eli Tilevich 01 May '25

01 May '25

Dear all, We are excited to share the May edition of the Computer Science Graduate Program Newsletter, a monthly publication of the Graduate Program Office. This edition will be the final release for this academic year. Best regards, Dr. T. ---------------------------------------------------------------------- Eli Tilevich, Ph.D. Professor and Associate Dept. Head for Graduate Studies Gilbert Place #4204 Dept. of Computer Science Virginia Tech Zoom: https://virginiatech.zoom.us/my/tilevich (540) 231-3475 http://people.cs.vt.edu/~tilevich Office hours: Monday/Tuesday 3 to 4pm in-person and on Zoom

1 0

FW: Fall 2025 Research Writing for Engineers Course
by Kinder-Potter, Sharon 01 May '25

01 May '25

From: Scarff, Kelly <kls16(a)vt.edu> Sent: Thursday, May 1, 2025 1:11 PM Hello, Would you please circulate this to your graduate students in case any are interested in taking the course? I'm a faculty member in the Mechanical Engineering department and am offering a graduate course in fall 2025: ME5984 Research Writing for Engineers. This course offers advanced writing lessons and aids students in learning the hallmark language and moves of common genres in engineering (journal articles, proposals, and literature reviews). It's a great course for graduate students who find themselves unsure of how to write the necessary documents for their degree (thesis/dissertation, literature review, research papers, etc.). Below is the timetable for the course, and attached is the course syllabus. Please Note for Students: This is a writing-intensive course. In order to provide students with quality one-on-one feedback on their writing, this course is hard-capped at 10 students. [cid:image001.png@01DBBA9A.2FA67690] Thank you, Kelly Scarff, PhD Collegiate Assistant Professor Director, Technical Communication Program Department of Mechanical Engineering 213A Goodwin Hall, Virginia Tech Blacksburg, VA 24061

1 0