+
+
+
+
+
+
+
+
+ 
+ A partnership between Georgia State University, Georgia Institute of Technology, and Emory University + +
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 
+ 3 postdocs, 14 PhD (5 of which recently defended), 5 MS, 4 undergrads, 1 volunteer
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ 
+
+
+
+ 
+
+
+
+ 
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ contact me with interesting projects in AI for Space
+
+
+
+
+ +
Machine Learning Capabilities of TReNDS Center:
+From Neuroimaging to Space +
+Sergey Plis
+Director of Machine Learning Core
+Associate Professor at CS Department
++
+ A partnership between Georgia State University, Georgia Institute of Technology, and Emory University + +
Tri-institutional center
+-
+
- Founded in 2019 +
- Employ >100 researchers +
- Including GRAs from GSU/GT/Emory +
- Publishing >120 journal papers a year +
- Publishing >100 conference papers a year +
- >15 PIs submitting proposals +
- ~50 proposals submitted a year +
- Annual burn rate ~$7M +
TReNDS guiding principles
+-
+
- From data to knowledge: we use data-driven approaches to inform theory and vice versa +
- Creation of new approaches and tools: we draw upon engineering and computer science principles to develop new algorithms to extract the maximal information possible from the available data +
- No data left unused: we work to effectively utilize available data to help point us in the right direction for future studies +
- Collaboration: we believe a highly collaborative & interdisciplinary environment is necessary to solve the hard problems we are facing today +
- Training: dissemination of knowledge is critical to advancing our understanding of the brain +
- Application: we desire to apply the approaches to address key challenges +
- Translation: We recognize the importance of tech transfer efforts +
Partnerships
+-
+
- GSU
+
-
+
- Psychology, CS, Neuroscience, Physics, Math, Business, Nursing, Public Health +
- Undergraduate/graduate training, programs for integrating research and education +
- Interdisciplinary imaging hub +
- GATech
+
-
+
- ECE, CSE, BME, Psychology +
- GTNeuro +
- GTRI +
- Emory
+
-
+
- Psychiatry, Neurology, Psychology, Radiology +
- Brain Health Personalized Medicine Institute +
- Markus Autism Center +
-
+
- Morehouse Medical +
- UT Dallas Center for BrainHealth & Brain Performance Institute + +
- The Mind Research Network +
- The Child Mind Institute +
- Rutgers University +
- Erasmus University +
- Martinos Center/MIT +
- Hartford Hospital +
- Yale University +
- Johns Hopkins Medical Institutions +
- University of California San Diego +
- And many more ... +
Outreach (funded by NFS)
+-
+
- Atlanta Science Festival Exploration Expo exhibitor +
- BRAINS! Center for Advanced Brain Imaging open house +
- Sponsored the Innovative Methodology and Diversity, Equity, & Inclusion awards at CABI Callosum +
- Sponsored the Neuroimaging in Aging Award at the GSU Aging Conference +
- Build-A-Brain at Drew Charter School +
- Virtual Neuroimaging info session with Q&A +
- MIND'S EYE: Student art competition at Drew Charter School +
- Atlanta Brain Awareness Campaign participants +
- Paideia School CABI field trip +
just my tiny team
+
+ 3 postdocs, 14 PhD (5 of which recently defended), 5 MS, 4 undergrads, 1 volunteer
+ Brain Imaging: MRI
+
+ Brain Imaging: fMRI
+ +Other data types
+-
+
- Diffusion tensor imaging +
- Magneto Encephalography +
- Electro Encephalography +
- Genomics data +
- Computerized tomography +
- and more +
Over 200,000 brain scans
+Data-driven methods
+machine learning
+artificial intelligence
+Independent Component analysis
+
+ Population Analysis
+
+
+ Causal Discovery
+Causal timescale $\ne$ Measurement timescale
++
+
+
+ What causal inferences can be made in this situation?
+
+ 
+
+ $\approx100ms$ +
+
+ + $\approx100ms$ +
+ 
+
+ $???$ +
+
+ + $???$ +
+ 
+
+ $\approx2s$ +
+
+
+
+ + $\approx2s$ +
Computational resources
+| Type | +Description | +
|---|---|
| + | 35 dedicated, hi-memory 28 core servers | +
| + | 36 high memory 64 core AMD CPU servers | +
| + | 40 v100 GPUs | +
| + | 32 A100 GPUs | +
| + | 3.8PB of local storage | +
Machine Learning Competencies
+-
+
- Telecom companies +
- Hospitals (mortality rate prediction) +
- Insurance claims in health informatics +
- Natural language processing +
- Embedded AI models +
- and more +
+