TRIPODS Winter School & Workshop- Stefanie Jegelka
Title- Learning and Generalization in Graph Neural Networks
Abstract- Graph Neural Networks (GNNs) have become a popular tool for learning representations of graph-structured inputs, with applications in computational chemistry, recommendation, pharmacy, reasoning, and many other areas. In this talk, I will show some recent results on learning with message-passing GNNs. For example, while several networks may be able to represent a task, some architectures learn it better than others. With a focus on algorithmic tasks, we hence look at inductive biases from the perspective of relating the architectural structure to the structure of the task, and study these theoretically and empirically. Next, we relate these results to predictions out of the training distribution, and contrast interpolation and extrapolation in GNNs. (Joint work with Keyulu Xu, Jingling Li, Mozhi Zhang, Simon S. Du, Ken-ichi Kawarabayashi, Vikas Garg and Tommi Jaakkola.)