6/18/2020 Aaron Seidlitz, Illinois CS 5 min read
Deepak Vasisht, an incoming Illinois CS professor, and two PhD graduates, Mengjia Yan and Silu Huang, developed dissertations that were among the best in their research areas.
Written by Aaron Seidlitz, Illinois CS
The dissertation is proof of a PhD candidate's capability in research and leads to the completion of their doctoral studies.
This one document is the culmination of several years of research and showcases the student’s ability to home in on an area of focus. Successful dissertations advance the state-of-the-art technology in their specific discipline.
Three PhD graduates associated with Illinois Computer Science recently produced dissertations at level worthy of international recognition:
- Deepak Vasisht, incoming Illinois CS assistant professor. Winner of the 2019 SIGCOMM Dissertation Award for his dissertation titled “Toward Realizing the Internet-of-Things Vision: In-body, Homes, and Farms”.
- Silu Huang (PhD CS ’19), current researcher with Microsoft Research. Honorable Mention for the 2020 ACM SIGMOD Jim Gray Dissertation Award. Her dissertation's title: “Effective Data Versioning for Collaborative Data Analytics”.
- Mengjia Yan (MS CS ’16, PhD CS ’19), current assistant professor with Massachusetts Technology Institute. Honorable Mention for the 2020 ACM SIGARCH/IEEE CS TCCA Outstanding Dissertation Award. Her dissertation's title: “Cache-based Side Channels: Modern Attacks and Defenses”.
“These accomplishments prove the researchers obtained results that pushed the boundaries of their field,” said Nancy M. Amato, Illinois CS department head. “The ingenuity of our students and expertise of our faculty drives our research enterprise. This combination produces a longstanding track record of research that transforms the way the world works.
“The recognition that these PhD graduates earned is sure to be just the first step on their journey to becoming global leaders through the impact of their work.”
Vasisht built two projects out of an innovative approach to the Internet-of-Things
As he approaches the beginning of his academic career at Illinois CS, as an assistant professor in January 2021, Vasisht said he was “overwhelmed” when he found out about his dissertation’s recognition by SIGCOMM.
The research he conducted regarding the Internet-of-Things (IoT) took six years and started under the guidance of his adviser, Professor Dina Katabi at the Massachusetts Institute of Technology. The two began investigating an indoor positioning system that could track people in their living environments for smart home applications. Vasisht then found further inspiration during an internship at Microsoft.
Throughout his dissertation work, Vasisht’s research has contributed to projects like FarmBeats – which is now publicly available as Microsoft Azure Farmbeats – and ReMix. He describes FarmBeats as “an end-to-end data-driven agricultural platform that can collect and analyze data from farms in remote rural areas at a very low cost compared to existing solutions.” ReMix came from an experience in conjunction with Massachusetts General Hospital, in which this technology was utilized like GPS to track in-body devices.
“The Internet-of-things has gained traction in recent years because of its ability to bridge our physical and digital worlds. I was thrilled to demonstrate that the techniques I built are useful through real-world IoT system implementations in farms, in-body and homes,” Vasisht said. “A lot of the credit goes to my adviser and my mentors over the years.”
Huang navigated the ups and downs to produce research that addresses issues in data science
The years it takes to produce research that results in a dissertation often takes the researchers in several different directions.
Huang’s experience was no different, but she ultimately found a niche that altered the way collaboration occurs within the realm of data science.
According to Huang’s research, there are three major issues with the way versioned datasets are currently managed in the wild. First, she found, there is “massive redundancy;” second there is a lack of true collaboration or sharing; and, third, the lack of querying capability means users need to write their own scripts.
The initial work in her first years of researching this provided the excitement Huang was hoping for. In the second year, her direction diverged a bit.
Still, she found the support she needed from adviser and Illinois CS adjunct professor Aditya Parameswaran as well as University of Chicago Computer Science professor Aaron Elmore. The combined sense of determination from this group and other collaborators, led to their solution –effective data versioning for collaborative data analytics, provided through a system they built called OrpheusDB.
“This experience has led me to a good starting point for my future work – I’m going to dream big and aim high,” Huang said. “Being recognized for the work means a lot and proves that I was fortunate to have had a very nice, encouraging and helpful adviser.”
Yan drilled down, found solutions to crucial security problems in modern processors
Starting her six-year research effort by “broadly exploring different hardware security problems,” Yan learned how much work was ahead.
It helped to have sound guidance from two Illinois CS professors – advisor Josep Torrellas and mentor Chris Fletcher. Their help propelled her forward. As time went on, she found a devotion regarding “micro-architecture side channel vulnerabilities” and the “defense mechanisms to mitigate those security attacks.”
Her solutions revolved around two notions.
First, she found new vulnerabilities and new opportunities for designing defense mechanisms in modern hierarchical caches, which results in papers published in top computer architecture and security conferences.
Second, her studies found a way to “effectively defend against transient execution attacks on cache hierarchies.” In a scheme she entitled “Invisible Speculation,” Yan detailed the first “robust hardware defense mechanism against transient cache-based side channel attacks for multiprocessors.” Her work applies to critical vulnerabilities often exploited by attacks like Meltdown and Spectre.
“My methods of completing the PhD thesis are no different than many other students – that is to be passionate and persistent. Additionally, Illinois CS has a great atmosphere and culture for doing research, which I benefited a lot from,” Yan said. “Earning this honorable mention distinction is a great honor, and it motivates me to continue doing high quality research.”