Name: Xinlian Liu
Name of the Company: Hood College
Business Area: Education
Position in the Company: Associate Professor, Fulbright Scholar in Data Science
INESC TEC Centre with which collaborated: High-Assurance Software Laboratory (HASLab)
Project managers (INESC TEC): Rui Oliveira and João Paulo
Title of the project: HPC Enabled Deep Learning of Biomedical Data
What is your professional background with INESC TEC?
I applied to be a Fulbright Scholar in Data Science which was hosted by the University of Minho and INSEC TEC. Rui Oliveira worked with me to develop the application and fortunately our proposal was selected for the award.
Briefly describe the main projects/collaborations you have carried out in collaboration with INESC TEC.
We worked together on a broad range of subjects in teaching, research, and administration in academia. My background was in HPC, but more recently focused on machine learning of extremely large data set. Handling such data-intensive applications presents both a challenge and an opportunity for data centers which traditionally were designed to process computation-intensive tasks. The High-Assurance Software Laboratory (HASLab) is a branch of the INESC TEC with many graduate students and faculty affiliated with the Lab. We collaborated to examine the I/O (input / output) behaviour of data centric applications and propose innovative solutions to meet new patterns of demands of resources.
What are the significant outcomes about these projects/collaborations that you can highlight?
Collaborating with the Berkeley Lab (LBL) where I am a visiting faculty, we implemented a workflow for deep learning of medical records to demonstrate a privacy-preserving computing platform that had been developed at the HASLab.
Collaborating with the Texas Advanced Computing Centre (TACC), we submitted a joint proposal to improve I/O performance for data-centric applications on HPC.
What is your opinion on your experience working with INESC TEC?
My experience working with Rui Oliveira, Joao Paulo, as well as other INESC TEC affiliated faculty and students, has been productive, fruitful, and joyful. We built bridges between research institutions in the US and Portugal. Hopefully, we may continue to make and to expand these connections.
What is the possibility of future collaborations with INESC TEC?
The expertise of high-performance computing at INESC TEC and my interests in machine learning/artificial intelligence naturally complements each other. We certainly hope to continue the collaboration, and we are actively working towards that.