2024-09-22: Looking Back at My PhD Journey

By -


I (Gavindya Jayawardena) began my PhD journey in Spring 2019 under the guidance of Dr. Sampath Jayarathna at Old Dominion University (ODU), immediately after completing my Bachelor’s degree. I joined the Neuro-Information Retrieval and Data Science (NIRDS) lab, a subgroup within the Web Science and Digital Libraries (WSDL) research lab in the Computer Science Department. Dr. Sampath Jayarathna provided me with the opportunity to engage in eye-tracking research, which rapidly became the central focus of my studies.

In collaboration with Dr. Anne Perrotti from ODU, I explored how eye-tracking measurements could be utilized to predict Attention-Deficit/Hyperactivity Disorder (ADHD) using machine learning algorithms. By extracting raw eye-tracking data, I created a feature set that enabled multiple models to achieve high prediction accuracy with tree-based classifiers. Additionally, I studied the performance of adolescents with ADHD during an audiovisual Speech-In-Noise task, analyzing cognitive load through eye-tracking measurements. This collaboration with Dr. Anne Perrott and Dr. Andrew Duchowski from Clemson University revealed that certain signal-to-noise ratios could increase cognitive load and impede speech processing.

My research on cognitive load and eye movement data led to the creation of RAEMAP (Real-Time Eye Movement Analysis Pipeline). RAEMAP is a novel software designed for real-time eye-tracking analysis, featuring algorithms that facilitate the computation of both basic and advanced gaze measures. This includes real-time assessments of pupil dilation for cognitive load, dynamic indicators of visual search behavior, and predictive analyses of gaze transitions. I conducted multiple evaluations across various modules over five years and demonstrated RAEMAP's effectiveness, particularly in assessing cognitive load and visual attention in real-time. My findings indicate that the real-time cognitive load measure can accurately identify shifts between baseline and cognitive load states, bridging the gap between traditional offline methods and the need for immediate insights into complex visual and cognitive processes.

I successfully defended my PhD on July 10, 2024, focusing on the development of RAEMAP. I have embedded the slides used in my defense presentation below. The dissertation will be available in ODU Digital Commons. 


I am immensely grateful to my dissertation committee members and mentors: Dr. Sampath Jayarathna, Dr. Anne Perrotti, Dr. Michael Nelson, Dr. Michele Weigle, Dr. Vikas Ashok, Dr. Andrew Duchowski, and Dr. Jian Wu, along with all others who have supported me throughout my PhD journey.

Since joining ODU, I also collaborated with various researchers on multiple projects, broadening my technological expertise. One notable collaboration involved developing a history-aware crawl scheduler for the academic web with Dr. Jian Wu and Dr. Alexander Nwala, resulting in a published paper. Additionally, I analyzed the reading patterns of novice researchers using eye-tracking metrics, which earned best poster awards at the ACM/IEEE Joint Conference on Digital Libraries for two consecutive years. I also collaborated with Dr. Bertrand Schneider at Harvard University’s Learning, Innovation and Technology Lab to explore the relationship between joint visual attention and joint mental effort in collaborative learning environments, analyzing eye-tracking datasets.

During the summers of 2021 and 2022, I worked as a Graduate Research Intern at Los Alamos National Laboratory (LANL), focusing on extracting features such as titles, authors, affiliations, abstracts, and keywords from scientific PDF documents using open-source, machine-learning tools like GROBID. This project aimed to process documents in real-time, and I developed a website to review the extracted metadata.

Throughout my PhD journey, I was honored to receive several scholarships, including the Dominion Graduate Scholarship (2019) and the Irwin B. Levinstein Scholarship (2020-2021), and I was recognized as one of the outstanding researchers in the Computer Science department.

Over the past few years, I’m grateful that my research has made significant contributions to the field of eye-tracking, resulting in multiple peer-reviewed publications. I am excited to further explore the intersection of neuro-physiological signals and human-computer interaction in my Postdoctoral Bullard Research Fellowship at the School of Information at the University of Texas at Austin, working with Dr. Jacek Gwizdka starting in Fall 2024.

To wrap up this blog post, I want to share a few lessons I learned:

Clear communication is key: It’s crucial to present your research clearly, both in writing and verbally. Take every opportunity to discuss your work. Practice makes perfect, and explaining your research in non-technical terms can enhance your own understanding. Personally, I struggled with presenting and dealt with imposter syndrome (who doesn’t?). Over time, I realized that most of the audience is there to learn from you, not to judge your work. I am also incredibly grateful to Dr. Nelson, Dr. Weigle, and Dr. Jayarathna for their feedback on my writing and presentations. Their insights helped me identify and correct many mistakes, and I will always cherish the lessons I learned from them.

Seek feedback: Embrace constructive criticism. I was fortunate to have supportive colleagues like Yasith, Bhanuka, Yasasi, Himarsha, Skanda, Kritika, Bathsheba, Alex, Shawn, and Sawood. Their valuable insights, engagement in user studies, and encouragement enriched my PhD journey, making it truly enjoyable and fruitful.

My advice to my younger self and to those just starting their PhD journey: 

At the beginning, it might seem like nothing is possible, but as time passes, you'll see the value of your efforts. Just don’t give up. It will all be worth it. Lastly, surround yourself with people who genuinely appreciate your hard work and provide constructive criticism in a supportive manner.

- Gavindya Jayawardena (@Gavindya2)



Comments

Post a Comment