Graph Attention Networks (GATs) are a type of neural network designed for graph-structured data. They were proposed by Petar Veličković and his colleagues in 2017.Graph Attention Network". GATs address the limitations of previous techniques based on graph convolutions or their approximations by using masked self-attentional layers. GATs' nodes are able to perform attention operations on their neighbor features, implicitly assigning different weights to different nodes without the need for costly matrix operations (such as inversion) or prior knowledge of the graph structure. This enables GATs to simultaneously address several key challenges of spectral-based graph neural networks and enables the model to be conveniently applied to inductive problems as well as traditional problems.