1: \begin{abstract}

2:     Implicit graph neural networks (GNNs) have emerged as a potential approach to enable GNNs to capture long-range dependencies effectively.

3:     %

4:     However, poorly designed implicit GNN layers can experience over-smoothing or may have limited adaptability to learn data geometry, potentially hindering their performance in graph learning problems. 

5:     %

6:     To address these issues, we introduce a geometric framework to design implicit graph diffusion layers based on a parameterized graph Laplacian operator. 

7:     %

8:     Our framework allows learning the geometry of vertex and edge spaces, as well as the graph gradient operator from data.

9:     %

10:     We further show how implicit GNN layers can be viewed as the fixed-point solution of a Dirichlet energy minimization problem and give conditions under which it may suffer from over-smoothing.

11:     %which can suffer from \textit{feature independence over-smoothing} and \textit{unlabelled over-smoothing}.

12:     %

13:     To overcome the over-smoothing problem, we design our implicit graph diffusion layer as the solution of a Dirichlet energy minimization problem with constraints on vertex features, enabling it to trade off smoothing with the preservation of node feature information.

14:     %capture long-range dependencies without encountering over-smoothing issues. 

15:     %

16:     With an appropriate hyperparameter set to be larger than the largest eigenvalue of the parameterized graph Laplacian, our framework guarantees a unique equilibrium and quick convergence.

17:     %

18:     Our models demonstrate better performance than leading implicit and explicit GNNs on benchmark datasets for node and graph classification tasks, with substantial accuracy improvements observed for some datasets.

19: %    certain heterophilic datasets, i.e., Chameleon and Squirrel.

20: \end{abstract}

21: