1 paper
Jan Hůla, David MojžÃÅ¡ek, JiÅà JaneÄek +2
In this contribution, we demonstrate that Graph Neural Networks and Transformers can learn to reason about geometric constraints. We train them to predict spatial position of point…