Given our two datasets, propose a general (or universal) graph representation of data that can capture
the UberMovement prediction task as a special case of edge-level prediction tasks, and
the ClimART prediction task as a special case of a node-level prediction tasks.
Demonstrate your proposed data representation and data processing method by
solving each task using two distinct instances of a GNN model
solving each task using one single instance of a GNN model that shares parameters among these tasks
Doing these two things may demonstrate that we are on track of developing a deep learning method for enhancing multi-task, multi-modal problem solving.
ArsamAryandoust
commented
1 year ago
Given our two datasets, propose a general (or universal) graph representation of data that can capture
Demonstrate your proposed data representation and data processing method by
Doing these two things may demonstrate that we are on track of developing a deep learning method for enhancing multi-task, multi-modal problem solving.