Can Graph Neural Networks Help Logic Inference? - 2019

Explore the potential of graph neural networks in bridging the gap between perceptual learning and logic inference in this insightful lecture by Le Song from Georgia Institute of Technology. Delve into the speaker's arguments on how graph neural networks can lead to new, efficient, and effective algorithms for challenging problems in inductive logic programming and lifted logic inference. Gain valuable insights into the future of AI as Song presents evidence suggesting that graph neural networks may be the key to advancing artificial intelligence to the next stage. Learn about the speaker's background, including his roles at Georgia Institute of Technology, Ant Financial, and his previous experiences at Carnegie Mellon University and Google. Discover Song's principal research directions in machine learning, kernel methods, deep learning, probabilistic graphical models, and relational data modeling, as well as his numerous awards and contributions to the field.