"The robust and reliable integration of machine learning techniques into existing scientific, medical and industrial technologies is one of the most important research challenges at the moment. Max’s research focuses on the potential of deep learning to dramatically improve modern imaging techniques, such as Magnetic Resonance Imaging (MRI) and microscopy. His work combines sophisticated mathematical tools - including optimization theory, approximation theory, compressed sensing and random matrix theory – to understand when and how deep learning can outperform existing start-of-the-art imaging techniques. Max’s cutting-edge research is characterized by both a high level of technical skill and a distinct mathematical flair. He is an absolute pleasure to work with."

- Dr. Ben Adcock

Meet More Students in Science

SEE MORE PROFILES

Curate your digital footprint

Want to be featured on our website? Complete our online submission form.

Submit your profile

Maksym Neyra-Nesterenko

October 14, 2021
Facebook
Twitter
LinkedIn
Reddit
SMS
Email
Copy

Applied and Computational Mathematics master's student in the Faculty of Science

Howdy! I am Maksym, or you can just call me Max. I have lived in Greater Vancouver for most of my life, and I am still here together with my wonderful immediate family and our pet cat. For leisurely activity, I do circuit training and mess around with configuring Linux desktop and server setups at home. Regarding school and research, I am very interested in the mathematics of machine learning. My interest stems from when I first heard about Google DeepMind achieving one of their major milestones. Here DeepMind’s Go board game program AlphaGo beat Lee Sedol, one of the highest-rated professional Go players in the world at the time. This happened back in 2016, where I was midway through my undergrad and have not yet heard the term “machine learning”. It seemed mystifying, nonetheless leaving me deeply curious. Since then, I have not ceased to be fascinated or curious about machine learning and its mathematics.

WHY DID YOU CHOOSE TO COME TO SFU?

I would say a mix of familiarity, proximity and research prospects. Not long ago, I completed my undergraduate degree in mathematics at SFU, so I know a bit about the math department and the students. Being near my friends, family, and the university has made things easier, especially during the COVID-19 pandemic. Lastly, I managed to connect with my current research supervisor, professor Ben Adcock, whose research interests and work are very close to my interests in machine learning and mathematics.

HOW WOULD YOU DESCRIBE YOUR RESEARCH OR YOUR PROGRAM TO A FAMILY MEMBER?

My research is in studying and developing accurate, robust and efficient artificial intelligence, specifically deep neural networks, for image reconstruction. Image reconstruction arises in everyday life, such as capturing photos with a phone camera, to advanced scientific tasks, such as medical imaging, electron microscopy and seismic imaging. Each of these involve acquiring measurements from a physical device and then constructing a digital image from these measurements. Many images possess some structure that allows for good image recovery using fewer measurements. This is advantageous for many reasons, such as less computation time and more resourceful design of scanners, cameras and other devices. Some tasks cannot afford to obtain many measurements, such as those in medical imaging, where more measurements lead to longer MRI scans, and expose patients to more radiation in X-ray scans. The current technology for image reconstruction is quite phenomenal, but with the advancements of artificial intelligence, can we take it farther? Deep neural networks are a powerful tool that has become integral to technology and scientific research. They can easily learn complex structures found in image-measurement data to perform image reconstruction. The big challenge is to meet all three desirable properties for image reconstruction, namely accuracy, robustness and efficiency. This is exactly what my research is about! Here accuracy means an accurate reconstruction of an image. Robustness refers to reliable reconstruction when encountering measurement noise or unseen data. Efficiency corresponds to fast reconstruction. Overall, the prospect of my research is to understand and study the limits in which artificial intelligence can push the boundaries in image reconstruction!

WHAT ARE YOU PARTICULARLY ENJOYING ABOUT YOUR STUDIES/RESEARCH AT SFU?

Unfortunately, the COVID-19 pandemic has limited the ability to interact and communicate with supervisors, instructors, fellow students and researchers. Despite this, I’ve been happy overall with the effort put in by everyone to facilitate communication remotely. The courses are intellectually stimulating, with some final exams substituted with course projects, and I have enjoyed participating in various workshops and giving contributed talks at conferences held online.

HAVE YOU BEEN THE RECIPIENT OF ANY MAJOR OR DONOR-FUNDED AWARDS? IF SO, PLEASE TELL US WHICH ONES AND A LITTLE ABOUT HOW THE AWARDS HAVE IMPACTED YOUR STUDIES AND/OR RESEARCH.

Yes, I have received two major scholarships during my M.Sc. program. I received the BC Graduate Scholarship (by nomination) and the NSERC CGS-Master’s (by application) in 2021. These were valued at $15,000 and $17,500, respectively, and have been of great financial assistance during the COVID-19 pandemic.

 

ADDITIONAL LINKS