Sahar Behpour has always had a desire for turning theories into applications and thoughts into actions.
Beginning her academic career as a physics Ph.D. candidate in 2017, she encountered research problems that were purely theoretical.
“How can I contribute more to application-based problems, or the data driven and universal approaches to solve the real problems in different domains such as healthcare, legal, finance, and social media?” Sahar asked herself. This was the genesis of her passion for data science. So, she made the degree switch from physics to data science.
After taking a machine learning course with Mark Albert, assistant professor of biomedical engineering, Sahar realized her strong interest in developing data driven, robust and universal artificial intelligence models that can work across different domains and datasets.
Working with her supervisor, Albert, who also has a background in computational neuroscience and computer science, Sahar began to investigate the sensory information processing in the brain— especially visual information so the computer will be able to learn things and make decisions in a humanlike manner.
While studying the visual information processing in the brain, also known as efficient coding, Sahar wondered, “how can we use similar processes when making artificial neuron networks?” From this inquiry, Sahar chose her dissertation topic: weight initialization of convolutional neural networking using unsupervised machine learning.
With each research discovery and focus, it leads Sahar to her ultimate goal of becoming a leader in the artificial intelligence field. Sahar hopes to create different infostructures to analyze data and models to provide a diverse set of services for the greater community.
“My goal is to not just make a profit for my business,” Sahar says, “but to be a resource for creating and improving science.”
Her dreams were allowed to flourish at UNT through the nurturing nature of the staff and faculty who encouraged her and assured her that she will be successful in this field.
“When I first walked into the information science department, Sahar called her family and said, ‘These people are so nice, I feel attached already!'”
As she gets closer to her goal, Sahar credits her accomplishments to her mentors and support system.
“Dr. Mark V. Albert taught me the fundamentals of machine learning and deep learning through his interactive and productive class and made me even more motivated to choose a dissertation topic in this direction. He showed me how to design and conduct a project from scratch while focusing on intuitions, writing succinctly, creating nice visualizations as part of any project, and finally delivering it at the right time. I have been greatly inspired by and benefited from his time management, professional communication, and leadership skills. Dr. Hawamdeh supported me throughout my whole Ph.D. journey. Dr. Paolo Grigolini constantly motivated me in research and making me more interested in science by sharing his in-depth knowledge every time I raised questions to him.“