Trainee Spotlight – Shi Dong

Shi Dong is a Research Assistant who has been working with PROTECT at Northeastern University under the mentorship of Data Core Leader Dave Kaeli for the last 5 years. He is currently pursuing a Ph.D. in Computer Engineering and researching “High Performance Computing and Big Data.” In his current role, Shi is responsible for doing research on cutting-edge high-performance computing systems, supporting emerging applications such as Deep Learning and Big Data, developing tools to facilitate the management and processing of Big Data applications, and performing big data processing and analytics using machine learning techniques. In the last year, Shi attended the 2018 and 2019 SRP Annual Meetings, the 2018 IEEE Big Data Conference, and the 2018 ACM ICPE. At these conferences, Shi presented posters and research papers, and won the best paper award at the 2018 ICPE for his work Characterizing Convolutional Neural Networks on GPUs.

Originally, Shi was drawn to PROTECT through a chance meeting with Professor Dave Kaeli in his Computer Architecture Course, and after becoming intrigued by the material subsequently contacted the professor to inquire about working in the lab. After securing the research assistant position, familiarizing himself with Professor Kaeli’s research, and becoming more involved in the lab Professor Kaeli offered Shi the opportunity to work in Core D of PROTECT. “After learning about the Center, I was excited about being part of this project because not only can I practice data analytics skills with a rich dataset, but also make contributions to the environmental health community. The idea of being able to help others inspires me and keeps me passionate.”

At the beginning of their mentor/mentee relationship, Professor Kaeli showed Shi the ropes and allowed him to participate in managing and maintaining the PROTECT database system as well as focusing on cleaning as much data as possible. Given the amount of data collected, they began to shift their focus to general data analytics leveraging Machine Learning techniques.

During his time with PROTECT, Shi has achieved many accomplishments and made various long-lasting contributions that will support the center for years to come. Shi developed a data processing framework and a toolset for facilitating the database management workflow, and conducted research for selecting important factors potentially contributing to the high preterm birth rate from data sets, one of the primary goals of the PROTECT Center. Through these accomplishments, Shi was able to gain skills in software engineering, data analytics, and machine learning which have in turn helped build a strong professional background for him to lean on post-graduation. When reflecting on this experience so far, Shi had this to say:

“The PROTECT Center is highly inter-disciplinary research oriented, bringing people from different backgrounds together for collaboration. This effort provides me unique opportunities to expand my knowledge base and sharpen my communication skills as well. The PROTECT Center has supported a number of learning opportunities and organized a rich set of activities during the SRP annual meeting, I have learned different methodologies targeting a broad range of problems in the environmental health field. Some of them provide me with valuable perspectives, sparking new ideas for me to improve my research work. Also, the training sessions offered by PROTECT help me hone my communication skills to better collaborate with other groups.”

“My favorite part of working with the PROTECT Center is also the opportunities that I can work with people from different backgrounds. Interestingly, one of the challenges also comes from this same benefit. Sharing my research work with people from diverse backgrounds is not always easy. I have to figure new ways to describe my work, while not scaring away people when they lack background knowledge. The trick to overcoming this challenge is to break a complex concept down into simple concepts that can be understood by all people.”

In the future, Shi plans to publish his compiled work on robust data analytics models within PROTECT which has strong performance in selecting essential features impacting the current high preterm birth rate. Shi is expected to graduate in July 2020 and is currently working on completing his thesis, wrapping up his current research, and pursuing a job in industry.

Shi Dong’s Publications & Posters

Publication:
A Hybrid Approach to Identifying Key Factors in Environmental Health Studies
An Efficient Data Management Framework for Puerto Rico Testsite for Exploring Contamination Threats (PROTECT)

Poster:
2019 Superfund Research Program Meeting
2018 Superfund Research Program Meeting
2018 Superfund Research Program Meeting