It's impossible to have a conversation about A.I. or machine learning without also talking about neural networks. The thing is, most of us think we know what it is, and some of us have an opinion on what machine learning is. But very few people actually know what a neural network is; why it matters? Do these networks evolve like the human brain? What are the ethical implications of building an infrastructure like that? Are they secure? My conversation today is with Sergey Plis, a professor of computer science at Georgia State University and the director of Machine Learning at the Center for Translational Research in Neuro Imaging and Data Science (TReNDS). He and his collaborators received funding recently from the NSF and the New Age to study casual connections in the brain. Obviously, Sergey likes to keep things busy. Please enjoy this incredible and fascinating conversation on this week's episode of The QTS Experience Podcast.
About Our Guest
As an Associate Professor of Computer Science and a Director of Machine Learning at the Center for Translational Research in Neuroimaging & Data Science, I like to keep things busy. Probably for that reason, I have three kids, two cats and a dog. I travel often, moved a few times recently, and have painted rooms and houses everywhere I go.
My research focuses on developing computational instruments that enable knowledge extraction from observational multimodal data collected at different temporal and spatial scales. I believe we can provide the neuroimaging community with a more robust, reliable understanding of directed connectivity in the brain.
We can’t just poke around in living human brain to see how it works. I and my collaborators are honored to have received $1.3 million from the National Science Foundation and the National Institutes of Health (NIH), to study causal connections in the brain. Here are only a few of my other accomplishments:
- Published one of the first demonstrations of the versatile potential of deep learning methods for the field in 2013 and since then my group developed a number of deep learning approaches to neuroimaging.
- Developed the theories in complex time series including a number of algorithms creating a new subarea in causal research.
- Developed efficient algorithms for matrix factorization.
- Enabled research on federated datasets with a focus on preserving privacy, inspiring a project for creating a framework to enable research on decentralized data.