2024-10-03: The End of a Long Journey: Reflecting on My PhD Experience

In the Spring of 2019, I embarked on a six-year journey toward earning my PhD in Computer Science. After completing my undergraduate degree in Computer Science and Engineering at the University of Moratuwa, Sri Lanka, in 2018, I applied to the Computer Science PhD program at Old Dominion University (ODU). I was fortunate to be accepted to the Neuro Information Retrieval and Data Science (NIRDS) Lab, which is part of ODU’s Web Science and Digital Libraries (WSDL) Research Group. Led by Dr. Sampath Jayarathna, the NIRDS lab focuses on applied research involving human subjects and multimodal biosignals.

Early in my PhD, before defining my research focus, I explored how to operate our lab's various biosignal acquisition devices and software (Eye Trackers, EEG Devices, and Wearable Health Sensors). This exploration allowed me to gain a deep understanding of the tools and technologies around human-subjects research, as well as their capabilities and limitations. A recurring issue that I faced was the difficulty of integrating data from multiple biosignal devices – without resorting to proprietary software. This issue got more complex when trying to integrate real-time biosensor data for multimodal analysis. While such integration was relatively straightforward when all devices were made by the same vendor, the reality was that no vendor provided biosignal hardware spanning all modalities.

This challenge became the cornerstone of my PhD research, driving me to find a better solution. Building on the insights I gained through exploration, I started developing an open-source multimodal stream processing framework, which we eventually named StreamingHub. My dissertation, titled “StreamingHub — A Real-Time Biosignal Processing Framework for Lab-Scale Experimentation,” focused on the design and development of this framework. The objective of StreamingHub was to allow easy integration of real-time biosignal data from different devices, while also supporting the replay of pre-recorded biosignal data as streams. To achieve this, I first proposed a JSON schema called DFDS (Data Flow Description Schema) to describe data sources, data collections, and the streams they generate (JCDL 2020 Paper). With DFDS in place, I developed a data integration and processing component called DataMux. This component allowed researchers to define (in Python) multimodal biosignal analysis pipelines that implicitly handle data access, without needing to code data access logic. To achieve this, DataMux relies on DFDS metadata, and assumes DFDS metadata already exists, which is often not true (yet). To address this limitation, I created an application called Curator, which offers researchers a simple, web-based interface to generate DFDS metadata for data sources and data collections.

These tools together formed a closed-loop ecosystem, designed to be open-source and modular, so that future researchers can easily extend the framework. For example, if a new biosignal device needs to be integrated, researchers can write a Python wrapper script (quite easily) to map the device's data access APIs into StreamingHub compatible format, enabling it to be used thereon for any multimodal biosignal analysis pipeline defined using StreamingHub.

Throughout the six years of my PhD journey, I conducted in-depth analyses of how researchers handle multimodal data streams, the software tools they rely on, and the open-source solutions available for integrating such data. Developing StreamingHub was not without challenges—one of the most significant was finding the right level of abstraction for the framework’s APIs so they could support a wide range of multimodal experiments without becoming overly complex. I successfully defended my PhD on July 09, 2024.

Ultimately, my research and the development of StreamingHub contributed to advancing the field of real-time biosignal processing, easing some of the technical challenges researchers face when working with multimodal data. Although the journey was filled with its share of obstacles, it was an immensely rewarding experience that not only shaped my academic career but also influenced the future of biosignal research tools. I am excited to be joining the Institute for Data Engineering and Science at Georgia Tech as a Research Scientist at the Center for Artificial Intelligence in Science and Engineering (ARTISAN), where I'll continue to build innovative tools and frameworks to streamline research workflows.

Before I conclude, I'd like to share a few things I wish my younger self had known early on during this journey.

Focus small, dream big: At the beginning, focusing on a small, well-defined problem felt almost insignificant - like solving it wouldn't have a broader impact by itself. What I didn't realize at the time was that tackling a focused issue creates a ripple effect. Each small solution contributes to a bigger picture, often in ways we can't immediately see. A PhD may not revolutionize your field overnight, but the momentum you build through incremental progress can ultimately shape larger outcomes with time.

Progress isn't always linear: There were stretches of time where I felt like I wasn’t moving forward. Research has a way of testing your patience; some breakthroughs happen in bursts, while other times you seem to be treading water. It’s important to trust the process and understand that even during slow periods, you’re still advancing. Setbacks and failures often lead to the most significant learning.

Learn to embrace uncertainty: Early in my PhD, I craved clarity—on my topic, on my next steps, on results. But uncertainty is part of the research journey. There will always be unknowns and shifting variables. The sooner you make peace with ambiguity and learn to navigate it, the more resilient you’ll become as a researcher.

Collaboration is key: PhD research can leave you feel isolated. Seeking input and collaboration from peers, mentors, and even those outside my field was invaluable. New perspectives can spark creative solutions and having a support system makes the ups and downs more manageable.

Balance depth with breadth: While deep focus on your topic is essential, don’t neglect the broader context of your field. Having a well-rounded understanding of adjacent areas can inform your own research and help you position your work within a larger framework. This makes it easier to communicate its relevance and impact.

Take care of your well-being: The intensity of PhD life can make it easy to lose sight of your physical and mental health. But burnout is real, and taking breaks is not a luxury—it’s a necessity. Prioritizing your well-being will ultimately help you stay focused, productive, and creative in the long run.

I owe a deep sense of gratitude to my advisor, whose ideas and guidance shaped my research and personal growth. I also want to thank my colleagues and fellow researchers for their invaluable feedback and friendship, which helped me stay grounded through the challenges. Finally, to my mentors—both formal and informal—your encouragement and insight were vital in helping me stay on course. This PhD would not have been possible without all of you.

-- Yasith Jayawardana (@yasithdev)