Machine Learning on Graphs (MLoG) Workshop
I am a Chief Scientist, Data Sciences in Advanced Computing, Mathematics and Data division at Pacific Northwest National Laboratory , and the co-director of our Computational and Theoretical Chemistry Institute . My current research focuses on scalable graph representation learning and neural-symbolic reasoning, with applications to chemistry, medical informatics and power grid. I have more than a decade's experience in developing artificial intelligence and data analytics systems that extract, learn and se
Nature is fond of geometry, and different molecular structures show unique geometric traits. However, training neural networks to predict properties of such geometric structures comes with a unique set of challenges. Our ongoing work focuses on accelerating the training of such neural networks on emerging AI accelerators, heterogeneous computing that integrates simulations and machine-learning on supercomputers, and creative combination of different machine-learning techniques to support wider range of chem
What do chemical compounds, power grid and Wikipedia have in common? All of them are a manifestation of diffrent entities coming togeher, something that is commonly modeled as graphs, or networks. A significant part of my career is dedicated to "connecting the dots": learning and building tools that often map into diverse fields in computer science: data mining techniques to extract structural patterns, searching a graph database for complex patterns, semantic reasoning using knowledge graphs, predictive an
Machine Learning on Graphs (MLoG) Workshop